PreprintPDF Available

YOLOv4: Optimal Speed and Accuracy of Object Detection

Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

There are a huge number of features which are said to improve Convolutional Neural Network (CNN) accuracy. Practical testing of combinations of such features on large datasets, and theoretical justification of the result, is required. Some features operate on certain models exclusively and for certain problems exclusively, or only for small-scale datasets; while some features, such as batch-normalization and residual-connections, are applicable to the majority of models, tasks, and datasets. We assume that such universal features include Weighted-Residual-Connections (WRC), Cross-Stage-Partial-connections (CSP), Cross mini-Batch Normalization (CmBN), Self-adversarial-training (SAT) and Mish-activation. We use new features: WRC, CSP, CmBN, SAT, Mish activation, Mosaic data augmentation, CmBN, DropBlock regularization, and CIoU loss, and combine some of them to achieve state-of-the-art results: 43.5% AP (65.7% AP50) for the MS COCO dataset at a realtime speed of ~65 FPS on Tesla V100. Source code is at https://github.com/AlexeyAB/darknet
Content may be subject to copyright.
YOLOv4: Optimal Speed and Accuracy of Object Detection
Alexey Bochkovskiy
alexeyab84@gmail.com
Chien-Yao Wang
Institute of Information Science
Academia Sinica, Taiwan
kinyiu@iis.sinica.edu.tw
Hong-Yuan Mark Liao
Institute of Information Science
Academia Sinica, Taiwan
liao@iis.sinica.edu.tw
Abstract
There are a huge number of features which are said to
improve Convolutional Neural Network (CNN) accuracy.
Practical testing of combinations of such features on large
datasets, and theoretical justification of the result, is re-
quired. Some features operate on certain models exclusively
and for certain problems exclusively, or only for small-scale
datasets; while some features, such as batch-normalization
and residual-connections, are applicable to the majority of
models, tasks, and datasets. We assume that such universal
features include Weighted-Residual-Connections (WRC),
Cross-Stage-Partial-connections (CSP), Cross mini-Batch
Normalization (CmBN), Self-adversarial-training (SAT)
and Mish-activation. We use new features: WRC, CSP,
CmBN, SAT, Mish activation, Mosaic data augmentation,
CmBN, DropBlock regularization, and CIoU loss, and com-
bine some of them to achieve state-of-the-art results: 43.5%
AP (65.7% AP50) for the MS COCO dataset at a real-
time speed of 65 FPS on Tesla V100. Source code is at
https://github.com/AlexeyAB/darknet.
1. Introduction
The majority of CNN-based object detectors are largely
applicable only for recommendation systems. For example,
searching for free parking spaces via urban video cameras
is executed by slow accurate models, whereas car collision
warning is related to fast inaccurate models. Improving
the real-time object detector accuracy enables using them
not only for hint generating recommendation systems, but
also for stand-alone process management and human input
reduction. Real-time object detector operation on conven-
tional Graphics Processing Units (GPU) allows their mass
usage at an affordable price. The most accurate modern
neural networks do not operate in real time and require large
number of GPUs for training with a large mini-batch-size.
We address such problems through creating a CNN that op-
erates in real-time on a conventional GPU, and for which
training requires only one conventional GPU.
Figure 1: Comparison of the proposed YOLOv4 and other
state-of-the-art object detectors. YOLOv4 runs twice faster
than EfficientDet with comparable performance. Improves
YOLOv3’s AP and FPS by 10% and 12%, respectively.
The main goal of this work is designing a fast operating
speed of an object detector in production systems and opti-
mization for parallel computations, rather than the low com-
putation volume theoretical indicator (BFLOP). We hope
that the designed object can be easily trained and used. For
example, anyone who uses a conventional GPU to train and
test can achieve real-time, high quality, and convincing ob-
ject detection results, as the YOLOv4 results shown in Fig-
ure 1. Our contributions are summarized as follows:
1. We develope an efficient and powerful object detection
model. It makes everyone can use a 1080 Ti or 2080 Ti
GPU to train a super fast and accurate object detector.
2. We verify the influence of state-of-the-art Bag-of-
Freebies and Bag-of-Specials methods of object detec-
tion during the detector training.
3. We modify state-of-the-art methods and make them
more effecient and suitable for single GPU training,
including CBN [89], PAN [49], SAM [85], etc.
1
arXiv:2004.10934v1 [cs.CV] 23 Apr 2020
Figure 2: Object detector.
2. Related work
2.1. Object detection models
A modern detector is usually composed of two parts,
a backbone which is pre-trained on ImageNet and a head
which is used to predict classes and bounding boxes of ob-
jects. For those detectors running on GPU platform, their
backbone could be VGG [68], ResNet [26], ResNeXt [86],
or DenseNet [30]. For those detectors running on CPU plat-
form, their backbone could be SqueezeNet [31], MobileNet
[28,66,27,74], or ShuffleNet [97,53]. As to the head part,
it is usually categorized into two kinds, i.e., one-stage object
detector and two-stage object detector. The most represen-
tative two-stage object detector is the R-CNN [19] series,
including fast R-CNN [18], faster R-CNN [64], R-FCN [9],
and Libra R-CNN [58]. It is also possible to make a two-
stage object detector an anchor-free object detector, such as
RepPoints [87]. As for one-stage object detector, the most
representative models are YOLO [61,62,63], SSD [50],
and RetinaNet [45]. In recent years, anchor-free one-stage
object detectors are developed. The detectors of this sort are
CenterNet [13], CornerNet [37,38], FCOS [78], etc. Object
detectors developed in recent years often insert some lay-
ers between backbone and head, and these layers are usu-
ally used to collect feature maps from different stages. We
can call it the neck of an object detector. Usually, a neck
is composed of several bottom-up paths and several top-
down paths. Networks equipped with this mechanism in-
clude Feature Pyramid Network (FPN) [44], Path Aggrega-
tion Network (PAN) [49], BiFPN [77], and NAS-FPN [17].
In addition to the above models, some researchers put their
emphasis on directly building a new backbone (DetNet [43],
DetNAS [7]) or a new whole model (SpineNet [12], HitDe-
tector [20]) for object detection.
To sum up, an ordinary object detector is composed of
several parts:
Input: Image, Patches, Image Pyramid
Backbones: VGG16 [68], ResNet-50 [26], SpineNet
[12], EfficientNet-B0/B7 [75], CSPResNeXt50 [81],
CSPDarknet53 [81]
Neck:
Additional blocks: SPP [25], ASPP [5], RFB
[47], SAM [85]
Path-aggregation blocks: FPN [44], PAN [49],
NAS-FPN [17], Fully-connected FPN, BiFPN
[77], ASFF [48], SFAM [98]
Heads::
Dense Prediction (one-stage):
RPN [64], SSD [50], YOLO [61], RetinaNet
[45] (anchor based)
CornerNet [37], CenterNet [13], MatrixNet
[60], FCOS [78] (anchor free)
Sparse Prediction (two-stage):
Faster R-CNN [64], R-FCN [9], Mask R-
CNN [23] (anchor based)
RepPoints [87] (anchor free)
2
2.2. Bag of freebies
Usually, a conventional object detector is trained off-
line. Therefore, researchers always like to take this advan-
tage and develop better training methods which can make
the object detector receive better accuracy without increas-
ing the inference cost. We call these methods that only
change the training strategy or only increase the training
cost as “bag of freebies.” What is often adopted by object
detection methods and meets the definition of bag of free-
bies is data augmentation. The purpose of data augmenta-
tion is to increase the variability of the input images, so that
the designed object detection model has higher robustness
to the images obtained from different environments. For
examples, photometric distortions and geometric distortions
are two commonly used data augmentation method and they
definitely benefit the object detection task. In dealing with
photometric distortion, we adjust the brightness, contrast,
hue, saturation, and noise of an image. For geometric dis-
tortion, we add random scaling, cropping, flipping, and ro-
tating.
The data augmentation methods mentioned above are all
pixel-wise adjustments, and all original pixel information in
the adjusted area is retained. In addition, some researchers
engaged in data augmentation put their emphasis on sim-
ulating object occlusion issues. They have achieved good
results in image classification and object detection. For ex-
ample, random erase [100] and CutOut [11] can randomly
select the rectangle region in an image and fill in a random
or complementary value of zero. As for hide-and-seek [69]
and grid mask [6], they randomly or evenly select multiple
rectangle regions in an image and replace them to all ze-
ros. If similar concepts are applied to feature maps, there
are DropOut [71], DropConnect [80], and DropBlock [16]
methods. In addition, some researchers have proposed the
methods of using multiple images together to perform data
augmentation. For example, MixUp [92] uses two images
to multiply and superimpose with different coefficient ra-
tios, and then adjusts the label with these superimposed ra-
tios. As for CutMix [91], it is to cover the cropped image
to rectangle region of other images, and adjusts the label
according to the size of the mix area. In addition to the
above mentioned methods, style transfer GAN [15] is also
used for data augmentation, and such usage can effectively
reduce the texture bias learned by CNN.
Different from the various approaches proposed above,
some other bag of freebies methods are dedicated to solving
the problem that the semantic distribution in the dataset may
have bias. In dealing with the problem of semantic distri-
bution bias, a very important issue is that there is a problem
of data imbalance between different classes, and this prob-
lem is often solved by hard negative example mining [72]
or online hard example mining [67] in two-stage object de-
tector. But the example mining method is not applicable
to one-stage object detector, because this kind of detector
belongs to the dense prediction architecture. Therefore Lin
et al. [45] proposed focal loss to deal with the problem
of data imbalance existing between various classes. An-
other very important issue is that it is difficult to express the
relationship of the degree of association between different
categories with the one-hot hard representation. This rep-
resentation scheme is often used when executing labeling.
The label smoothing proposed in [73] is to convert hard la-
bel into soft label for training, which can make model more
robust. In order to obtain a better soft label, Islam et al. [33]
introduced the concept of knowledge distillation to design
the label refinement network.
The last bag of freebies is the objective function of
Bounding Box (BBox) regression. The traditional object
detector usually uses Mean Square Error (MSE) to di-
rectly perform regression on the center point coordinates
and height and width of the BBox, i.e., {xcenter ,ycenter,
w,h}, or the upper left point and the lower right point,
i.e., {xtop lef t,ytop lef t,xbottom right,ybottom right }. As
for anchor-based method, it is to estimate the correspond-
ing offset, for example {xcenter of f set,ycenter of fset ,
woffset ,hoff set}and {xtop lef t of fset ,ytop lef t off set,
xbottom right of f set,ybottom right of f set}. However, to di-
rectly estimate the coordinate values of each point of the
BBox is to treat these points as independent variables, but
in fact does not consider the integrity of the object itself. In
order to make this issue processed better, some researchers
recently proposed IoU loss [90], which puts the coverage of
predicted BBox area and ground truth BBox area into con-
sideration. The IoU loss computing process will trigger the
calculation of the four coordinate points of the BBox by ex-
ecuting IoU with the ground truth, and then connecting the
generated results into a whole code. Because IoU is a scale
invariant representation, it can solve the problem that when
traditional methods calculate the l1or l2loss of {x,y,w,
h}, the loss will increase with the scale. Recently, some
researchers have continued to improve IoU loss. For exam-
ple, GIoU loss [65] is to include the shape and orientation
of object in addition to the coverage area. They proposed to
find the smallest area BBox that can simultaneously cover
the predicted BBox and ground truth BBox, and use this
BBox as the denominator to replace the denominator origi-
nally used in IoU loss. As for DIoU loss [99], it additionally
considers the distance of the center of an object, and CIoU
loss [99], on the other hand simultaneously considers the
overlapping area, the distance between center points, and
the aspect ratio. CIoU can achieve better convergence speed
and accuracy on the BBox regression problem.
3
2.3. Bag of specials
For those plugin modules and post-processing methods
that only increase the inference cost by a small amount
but can significantly improve the accuracy of object detec-
tion, we call them “bag of specials”. Generally speaking,
these plugin modules are for enhancing certain attributes in
a model, such as enlarging receptive field, introducing at-
tention mechanism, or strengthening feature integration ca-
pability, etc., and post-processing is a method for screening
model prediction results.
Common modules that can be used to enhance recep-
tive field are SPP [25], ASPP [5], and RFB [47]. The
SPP module was originated from Spatial Pyramid Match-
ing (SPM) [39], and SPMs original method was to split fea-
ture map into several d×dequal blocks, where dcan be
{1,2,3, ...}, thus forming spatial pyramid, and then extract-
ing bag-of-word features. SPP integrates SPM into CNN
and use max-pooling operation instead of bag-of-word op-
eration. Since the SPP module proposed by He et al. [25]
will output one dimensional feature vector, it is infeasible to
be applied in Fully Convolutional Network (FCN). Thus in
the design of YOLOv3 [63], Redmon and Farhadi improve
SPP module to the concatenation of max-pooling outputs
with kernel size k×k, where k={1,5,9,13}, and stride
equals to 1. Under this design, a relatively large k×kmax-
pooling effectively increase the receptive field of backbone
feature. After adding the improved version of SPP module,
YOLOv3-608 upgrades AP50 by 2.7% on the MS COCO
object detection task at the cost of 0.5% extra computation.
The difference in operation between ASPP [5] module and
improved SPP module is mainly from the original k×kker-
nel size, max-pooling of stride equals to 1 to several 3×3
kernel size, dilated ratio equals to k, and stride equals to 1
in dilated convolution operation. RFB module is to use sev-
eral dilated convolutions of k×kkernel, dilated ratio equals
to k, and stride equals to 1 to obtain a more comprehensive
spatial coverage than ASPP. RFB [47] only costs 7% extra
inference time to increase the AP50 of SSD on MS COCO
by 5.7%.
The attention module that is often used in object detec-
tion is mainly divided into channel-wise attention and point-
wise attention, and the representatives of these two atten-
tion models are Squeeze-and-Excitation (SE) [29] and Spa-
tial Attention Module (SAM) [85], respectively. Although
SE module can improve the power of ResNet50 in the Im-
ageNet image classification task 1% top-1 accuracy at the
cost of only increasing the computational effort by 2%, but
on a GPU usually it will increase the inference time by
about 10%, so it is more appropriate to be used in mobile
devices. But for SAM, it only needs to pay 0.1% extra cal-
culation and it can improve ResNet50-SE 0.5% top-1 accu-
racy on the ImageNet image classification task. Best of all,
it does not affect the speed of inference on the GPU at all.
In terms of feature integration, the early practice is to use
skip connection [51] or hyper-column [22] to integrate low-
level physical feature to high-level semantic feature. Since
multi-scale prediction methods such as FPN have become
popular, many lightweight modules that integrate different
feature pyramid have been proposed. The modules of this
sort include SFAM [98], ASFF [48], and BiFPN [77]. The
main idea of SFAM is to use SE module to execute channel-
wise level re-weighting on multi-scale concatenated feature
maps. As for ASFF, it uses softmax as point-wise level re-
weighting and then adds feature maps of different scales.
In BiFPN, the multi-input weighted residual connections is
proposed to execute scale-wise level re-weighting, and then
add feature maps of different scales.
In the research of deep learning, some people put their
focus on searching for good activation function. A good
activation function can make the gradient more efficiently
propagated, and at the same time it will not cause too
much extra computational cost. In 2010, Nair and Hin-
ton [56] propose ReLU to substantially solve the gradient
vanish problem which is frequently encountered in tradi-
tional tanh and sigmoid activation function. Subsequently,
LReLU [54], PReLU [24], ReLU6 [28], Scaled Exponential
Linear Unit (SELU) [35], Swish [59], hard-Swish [27], and
Mish [55], etc., which are also used to solve the gradient
vanish problem, have been proposed. The main purpose of
LReLU and PReLU is to solve the problem that the gradi-
ent of ReLU is zero when the output is less than zero. As
for ReLU6 and hard-Swish, they are specially designed for
quantization networks. For self-normalizing a neural net-
work, the SELU activation function is proposed to satisfy
the goal. One thing to be noted is that both Swish and Mish
are continuously differentiable activation function.
The post-processing method commonly used in deep-
learning-based object detection is NMS, which can be used
to filter those BBoxes that badly predict the same ob-
ject, and only retain the candidate BBoxes with higher re-
sponse. The way NMS tries to improve is consistent with
the method of optimizing an objective function. The orig-
inal method proposed by NMS does not consider the con-
text information, so Girshick et al. [19] added classification
confidence score in R-CNN as a reference, and according to
the order of confidence score, greedy NMS was performed
in the order of high score to low score. As for soft NMS [1],
it considers the problem that the occlusion of an object may
cause the degradation of confidence score in greedy NMS
with IoU score. The DIoU NMS [99] developers way of
thinking is to add the information of the center point dis-
tance to the BBox screening process on the basis of soft
NMS. It is worth mentioning that, since none of above post-
processing methods directly refer to the captured image fea-
tures, post-processing is no longer required in the subse-
quent development of an anchor-free method.
4
Table 1: Parameters of neural networks for image classification.
Backbone model Input network
resolution
Receptive
field size Parameters
Average size
of layer output
(WxHxC)
BFLOPs
(512x512 network resolution)
FPS
(GPU RTX 2070)
CSPResNext50 512x512 425x425 20.6 M 1058 K 31 (15.5 FMA) 62
CSPDarknet53 512x512 725x725 27.6 M 950 K 52 (26.0 FMA) 66
EfficientNet-B3 (ours) 512x512 1311x1311 12.0 M 668 K 11 (5.5 FMA) 26
3. Methodology
The basic aim is fast operating speed of neural network,
in production systems and optimization for parallel compu-
tations, rather than the low computation volume theoreti-
cal indicator (BFLOP). We present two options of real-time
neural networks:
For GPU we use a small number of groups (1 - 8) in
convolutional layers: CSPResNeXt50 / CSPDarknet53
For VPU - we use grouped-convolution, but we re-
frain from using Squeeze-and-excitement (SE) blocks
- specifically this includes the following models:
EfficientNet-lite / MixNet [76] / GhostNet [21] / Mo-
bileNetV3
3.1. Selection of architecture
Our objective is to find the optimal balance among the in-
put network resolution, the convolutional layer number, the
parameter number (filter size2* filters * channel / groups),
and the number of layer outputs (filters). For instance, our
numerous studies demonstrate that the CSPResNext50 is
considerably better compared to CSPDarknet53 in terms
of object classification on the ILSVRC2012 (ImageNet)
dataset [10]. However, conversely, the CSPDarknet53 is
better compared to CSPResNext50 in terms of detecting ob-
jects on the MS COCO dataset [46].
The next objective is to select additional blocks for in-
creasing the receptive field and the best method of parame-
ter aggregation from different backbone levels for different
detector levels: e.g. FPN, PAN, ASFF, BiFPN.
A reference model which is optimal for classification is
not always optimal for a detector. In contrast to the classi-
fier, the detector requires the following:
Higher input network size (resolution) – for detecting
multiple small-sized objects
More layers – for a higher receptive field to cover the
increased size of input network
More parameters – for greater capacity of a model to
detect multiple objects of different sizes in a single im-
age
Hypothetically speaking, we can assume that a model
with a larger receptive field size (with a larger number of
convolutional layers 3×3) and a larger number of parame-
ters should be selected as the backbone. Table 1shows the
information of CSPResNeXt50, CSPDarknet53, and Effi-
cientNet B3. The CSPResNext50 contains only 16 convo-
lutional layers 3×3, a 425 ×425 receptive field and 20.6
M parameters, while CSPDarknet53 contains 29 convolu-
tional layers 3×3, a 725 ×725 receptive field and 27.6
M parameters. This theoretical justification, together with
our numerous experiments, show that CSPDarknet53 neu-
ral network is the optimal model of the two as the backbone
for a detector.
The influence of the receptive field with different sizes is
summarized as follows:
Up to the object size - allows viewing the entire object
Up to network size - allows viewing the context around
the object
Exceeding the network size - increases the number of
connections between the image point and the final ac-
tivation
We add the SPP block over the CSPDarknet53, since it
significantly increases the receptive field, separates out the
most significant context features and causes almost no re-
duction of the network operation speed. We use PANet as
the method of parameter aggregation from different back-
bone levels for different detector levels, instead of the FPN
used in YOLOv3.
Finally, we choose CSPDarknet53 backbone, SPP addi-
tional module, PANet path-aggregation neck, and YOLOv3
(anchor based) head as the architecture of YOLOv4.
In the future we plan to expand significantly the content
of Bag of Freebies (BoF) for the detector, which theoreti-
cally can address some problems and increase the detector
accuracy, and sequentially check the influence of each fea-
ture in an experimental fashion.
We do not use Cross-GPU Batch Normalization (CGBN
or SyncBN) or expensive specialized devices. This al-
lows anyone to reproduce our state-of-the-art outcomes on
a conventional graphic processor e.g. GTX 1080Ti or RTX
2080Ti.
5
3.2. Selection of BoF and BoS
For improving the object detection training, a CNN usu-
ally uses the following:
Activations: ReLU, leaky-ReLU, parametric-ReLU,
ReLU6, SELU, Swish, or Mish
Bounding box regression loss: MSE, IoU, GIoU,
CIoU, DIoU
Data augmentation: CutOut, MixUp, CutMix
Regularization method: DropOut, DropPath [36],
Spatial DropOut [79], or DropBlock
Normalization of the network activations by their
mean and variance: Batch Normalization (BN) [32],
Cross-GPU Batch Normalization (CGBN or SyncBN)
[93], Filter Response Normalization (FRN) [70], or
Cross-Iteration Batch Normalization (CBN) [89]
Skip-connections: Residual connections, Weighted
residual connections, Multi-input weighted residual
connections, or Cross stage partial connections (CSP)
As for training activation function, since PReLU and
SELU are more difficult to train, and ReLU6 is specifically
designed for quantization network, we therefore remove the
above activation functions from the candidate list. In the
method of reqularization, the people who published Drop-
Block have compared their method with other methods in
detail, and their regularization method has won a lot. There-
fore, we did not hesitate to choose DropBlock as our reg-
ularization method. As for the selection of normalization
method, since we focus on a training strategy that uses only
one GPU, syncBN is not considered.
3.3. Additional improvements
In order to make the designed detector more suitable for
training on single GPU, we made additional design and im-
provement as follows:
We introduce a new method of data augmentation Mo-
saic, and Self-Adversarial Training (SAT)
We select optimal hyper-parameters while applying
genetic algorithms
We modify some exsiting methods to make our design
suitble for efficient training and detection - modified
SAM, modified PAN, and Cross mini-Batch Normal-
ization (CmBN)
Mosaic represents a new data augmentation method that
mixes 4 training images. Thus 4 different contexts are
Figure 3: Mosaic represents a new method of data augmen-
tation.
mixed, while CutMix mixes only 2 input images. This al-
lows detection of objects outside their normal context. In
addition, batch normalization calculates activation statistics
from 4 different images on each layer. This significantly
reduces the need for a large mini-batch size.
Self-Adversarial Training (SAT) also represents a new
data augmentation technique that operates in 2 forward
backward stages. In the 1st stage the neural network alters
the original image instead of the network weights. In this
way the neural network executes an adversarial attack on it-
self, altering the original image to create the deception that
there is no desired object on the image. In the 2nd stage, the
neural network is trained to detect an object on this modified
image in the normal way.
Figure 4: Cross mini-Batch Normalization.
CmBN represents a CBN modified version, as shown
in Figure 4, defined as Cross mini-Batch Normalization
(CmBN). This collects statistics only between mini-batches
within a single batch.
We modify SAM from spatial-wise attention to point-
wise attention, and replace shortcut connection of PAN to
concatenation, as shown in Figure 5and Figure 6, respec-
tively.
6
Figure 5: Modified SAM.
Figure 6: Modified PAN.
3.4. YOLOv4
In this section, we shall elaborate the details of YOLOv4.
YOLOv4 consists of:
Backbone: CSPDarknet53 [81]
Neck: SPP [25], PAN [49]
Head: YOLOv3 [63]
YOLO v4 uses:
Bag of Freebies (BoF) for backbone: CutMix and
Mosaic data augmentation, DropBlock regularization,
Class label smoothing
Bag of Specials (BoS) for backbone: Mish activa-
tion, Cross-stage partial connections (CSP), Multi-
input weighted residual connections (MiWRC)
Bag of Freebies (BoF) for detector: CIoU-loss,
CmBN, DropBlock regularization, Mosaic data aug-
mentation, Self-Adversarial Training, Eliminate grid
sensitivity, Using multiple anchors for a single ground
truth, Cosine annealing scheduler [52], Optimal hyper-
parameters, Random training shapes
Bag of Specials (BoS) for detector: Mish activation,
SPP-block, SAM-block, PAN path-aggregation block,
DIoU-NMS
4. Experiments
We test the influence of different training improve-
ment techniques on accuracy of the classifier on ImageNet
(ILSVRC 2012 val) dataset, and then on the accuracy of the
detector on MS COCO (test-dev 2017) dataset.
4.1. Experimental setup
In ImageNet image classification experiments, the de-
fault hyper-parameters are as follows: the training steps is
8,000,000; the batch size and the mini-batch size are 128
and 32, respectively; the polynomial decay learning rate
scheduling strategy is adopted with initial learning rate 0.1;
the warm-up steps is 1000; the momentum and weight de-
cay are respectively set as 0.9 and 0.005. All of our BoS
experiments use the same hyper-parameter as the default
setting, and in the BoF experiments, we add an additional
50% training steps. In the BoF experiments, we verify
MixUp, CutMix, Mosaic, Bluring data augmentation, and
label smoothing regularization methods. In the BoS experi-
ments, we compared the effects of LReLU, Swish, and Mish
activation function. All experiments are trained with a 1080
Ti or 2080 Ti GPU.
In MS COCO object detection experiments, the de-
fault hyper-parameters are as follows: the training steps is
500,500; the step decay learning rate scheduling strategy is
adopted with initial learning rate 0.01 and multiply with a
factor 0.1 at the 400,000 steps and the 450,000 steps, re-
spectively; The momentum and weight decay are respec-
tively set as 0.9 and 0.0005. All architectures use a sin-
gle GPU to execute multi-scale training in the batch size
of 64 while mini-batch size is 8 or 4 depend on the ar-
chitectures and GPU memory limitation. Except for us-
ing genetic algorithm for hyper-parameter search experi-
ments, all other experiments use default setting. Genetic
algorithm used YOLOv3-SPP to train with GIoU loss and
search 300 epochs for min-val 5k sets. We adopt searched
learning rate 0.00261, momentum 0.949, IoU threshold for
assigning ground truth 0.213, and loss normalizer 0.07 for
genetic algorithm experiments. We have verified a large
number of BoF, including grid sensitivity elimination, mo-
saic data augmentation, IoU threshold, genetic algorithm,
class label smoothing, cross mini-batch normalization, self-
adversarial training, cosine annealing scheduler, dynamic
mini-batch size, DropBlock, Optimized Anchors, different
kind of IoU losses. We also conduct experiments on various
BoS, including Mish, SPP, SAM, RFB, BiFPN, and Gaus-
sian YOLO [8]. For all experiments, we only use one GPU
for training, so techniques such as syncBN that optimizes
multiple GPUs are not used.
7
4.2. Influence of different features on Classifier
training
First, we study the influence of different features on
classifier training; specifically, the influence of Class la-
bel smoothing, the influence of different data augmentation
techniques, bilateral blurring, MixUp, CutMix and Mosaic,
as shown in Fugure 7, and the influence of different activa-
tions, such as Leaky-ReLU (by default), Swish, and Mish.
Figure 7: Various method of data augmentation.
In our experiments, as illustrated in Table 2, the clas-
sifier’s accuracy is improved by introducing the features
such as: CutMix and Mosaic data augmentation, Class la-
bel smoothing, and Mish activation. As a result, our BoF-
backbone (Bag of Freebies) for classifier training includes
the following: CutMix and Mosaic data augmentation and
Class label smoothing. In addition we use Mish activation
as a complementary option, as shown in Table 2and Table
3.
Table 2: Influence of BoF and Mish on the CSPResNeXt-50 clas-
sifier accuracy.
MixUp CutMix Mosaic Bluring Label
Smoothing Swish Mish Top-1 Top-5
77.9% 94.0%
X77.2% 94.0%
X78.0% 94.3%
X78.1% 94.5%
X77.5% 93.8%
X78.1% 94.4%
X64.5% 86.0%
X78.9% 94.5%
X X X 78.5% 94.8%
X X X X 79.8% 95.2%
Table 3: Influence of BoF and Mish on the CSPDarknet-53 classi-
fier accuracy.
MixUp CutMix Mosaic Bluring Label
Smoothing Swish Mish Top-1 Top-5
77.2% 93.6%
X X X 77.8% 94.4%
X X X X 78.7% 94.8%
4.3. Influence of different features on Detector
training
Further study concerns the influence of different Bag-of-
Freebies (BoF-detector) on the detector training accuracy,
as shown in Table 4. We significantly expand the BoF list
through studying different features that increase the detector
accuracy without affecting FPS:
S: Eliminate grid sensitivity the equation bx=σ(tx)+
cx, by=σ(ty)+cy, where cxand cyare always whole
numbers, is used in YOLOv3 for evaluating the ob-
ject coordinates, therefore, extremely high txabsolute
values are required for the bxvalue approaching the
cxor cx+ 1 values. We solve this problem through
multiplying the sigmoid by a factor exceeding 1.0, so
eliminating the effect of grid on which the object is
undetectable.
M: Mosaic data augmentation - using the 4-image mo-
saic during training instead of single image
IT: IoU threshold - using multiple anchors for a single
ground truth IoU (truth, anchor) >IoU threshold
GA: Genetic algorithms - using genetic algorithms for
selecting the optimal hyperparameters during network
training on the first 10% of time periods
LS: Class label smoothing - using class label smooth-
ing for sigmoid activation
CBN: CmBN - using Cross mini-Batch Normalization
for collecting statistics inside the entire batch, instead
of collecting statistics inside a single mini-batch
CA: Cosine annealing scheduler - altering the learning
rate during sinusoid training
DM: Dynamic mini-batch size - automatic increase of
mini-batch size during small resolution training by us-
ing Random training shapes
OA: Optimized Anchors - using the optimized anchors
for training with the 512x512 network resolution
GIoU, CIoU, DIoU, MSE - using different loss algo-
rithms for bounded box regression
Further study concerns the influence of different Bag-
of-Specials (BoS-detector) on the detector training accu-
racy, including PAN, RFB, SAM, Gaussian YOLO (G), and
ASFF, as shown in Table 5. In our experiments, the detector
gets best performance when using SPP, PAN, and SAM.
8
Table 4: Ablation Studies of Bag-of-Freebies. (CSPResNeXt50-PANet-SPP, 512x512).
S M IT GA LS CBN CA DM OA loss AP AP50 AP75
MSE 38.0% 60.0% 40.8%
XMSE 37.7% 59.9% 40.5%
XMSE 39.1% 61.8% 42.0%
XMSE 36.9% 59.7% 39.4%
XMSE 38.9% 61.7% 41.9%
XMSE 33.0% 55.4% 35.4%
XMSE 38.4% 60.7% 41.3%
XMSE 38.7% 60.7% 41.9%
XMSE 35.3% 57.2% 38.0%
XGIoU 39.4% 59.4% 42.5%
XDIoU 39.1% 58.8% 42.1%
XCIoU 39.6% 59.2% 42.6%
X X X X CIoU 41.5% 64.0% 44.8%
X X X CIoU 36.1% 56.5% 38.4%
X X X X X MSE 40.3% 64.0% 43.1%
X X X X X GIoU 42.4% 64.4% 45.9%
X X X X X CIoU 42.4% 64.4% 45.9%
Table 5: Ablation Studies of Bag-of-Specials. (Size 512x512).
Model AP AP50 AP75
CSPResNeXt50-PANet-SPP 42.4% 64.4% 45.9%
CSPResNeXt50-PANet-SPP-RFB 41.8% 62.7% 45.1%
CSPResNeXt50-PANet-SPP-SAM 42.7% 64.6% 46.3%
CSPResNeXt50-PANet-SPP-SAM-G 41.6% 62.7% 45.0%
CSPResNeXt50-PANet-SPP-ASFF-RFB 41.1% 62.6% 44.4%
4.4. Influence of different backbones and pre-
trained weightings on Detector training
Further on we study the influence of different backbone
models on the detector accuracy, as shown in Table 6. We
notice that the model characterized with the best classifica-
tion accuracy is not always the best in terms of the detector
accuracy.
First, although classification accuracy of CSPResNeXt-
50 models trained with different features is higher compared
to CSPDarknet53 models, the CSPDarknet53 model shows
higher accuracy in terms of object detection.
Second, using BoF and Mish for the CSPResNeXt50
classifier training increases its classification accuracy, but
further application of these pre-trained weightings for de-
tector training reduces the detector accuracy. However, us-
ing BoF and Mish for the CSPDarknet53 classifier training
increases the accuracy of both the classifier and the detector
which uses this classifier pre-trained weightings. The net
result is that backbone CSPDarknet53 is more suitable for
the detector than for CSPResNeXt50.
We observe that the CSPDarknet53 model demonstrates
a greater ability to increase the detector accuracy owing to
various improvements.
Table 6: Using different classifier pre-trained weightings for de-
tector training (all other training parameters are similar in all mod-
els) .
Model (with optimal setting) Size AP AP50 AP75
CSPResNeXt50-PANet-SPP 512x512 42.4 64.4 45.9
CSPResNeXt50-PANet-SPP
(BoF-backbone) 512x512 42.3 64.3 45.7
CSPResNeXt50-PANet-SPP
(BoF-backbone + Mish) 512x512 42.3 64.2 45.8
CSPDarknet53-PANet-SPP
(BoF-backbone) 512x512 42.4 64.5 46.0
CSPDarknet53-PANet-SPP
(BoF-backbone + Mish) 512x512 43.0 64.9 46.5
4.5. Influence of different mini-batch size on Detec-
tor training
Finally, we analyze the results obtained with models
trained with different mini-batch sizes, and the results are
shown in Table 7. From the results shown in Table 7, we
found that after adding BoF and BoS training strategies, the
mini-batch size has almost no effect on the detector’s per-
formance. This result shows that after the introduction of
BoF and BoS, it is no longer necessary to use expensive
GPUs for training. In other words, anyone can use only a
conventional GPU to train an excellent detector.
Table 7: Using different mini-batch size for detector training.
Model (without OA) Size AP AP50 AP75
CSPResNeXt50-PANet-SPP
(without BoF/BoS, mini-batch 4) 608 37.1 59.2 39.9
CSPResNeXt50-PANet-SPP
(without BoF/BoS, mini-batch 8) 608 38.4 60.6 41.6
CSPDarknet53-PANet-SPP
(with BoF/BoS, mini-batch 4) 512 41.6 64.1 45.0
CSPDarknet53-PANet-SPP
(with BoF/BoS, mini-batch 8) 512 41.7 64.2 45.2
9
Figure 8: Comparison of the speed and accuracy of different object detectors. (Some articles stated the FPS of their detectors
for only one of the GPUs: Maxwell/Pascal/Volta)
5. Results
Comparison of the results obtained with other state-
of-the-art object detectors are shown in Figure 8. Our
YOLOv4 are located on the Pareto optimality curve and are
superior to the fastest and most accurate detectors in terms
of both speed and accuracy.
Since different methods use GPUs of different architec-
tures for inference time verification, we operate YOLOv4
on commonly adopted GPUs of Maxwell, Pascal, and Volta
architectures, and compare them with other state-of-the-art
methods. Table 8lists the frame rate comparison results of
using Maxwell GPU, and it can be GTX Titan X (Maxwell)
or Tesla M40 GPU. Table 9lists the frame rate comparison
results of using Pascal GPU, and it can be Titan X (Pascal),
Titan Xp, GTX 1080 Ti, or Tesla P100 GPU. As for Table
10, it lists the frame rate comparison results of using Volta
GPU, and it can be Titan Volta or Tesla V100 GPU.
6. Conclusions
We offer a state-of-the-art detector which is faster (FPS)
and more accurate (MS COCO AP50...95 and AP50) than
all available alternative detectors. The detector described
can be trained and used on a conventional GPU with 8-16
GB-VRAM this makes its broad use possible. The original
concept of one-stage anchor-based detectors has proven its
viability. We have verified a large number of features, and
selected for use such of them for improving the accuracy of
both the classifier and the detector. These features can be
used as best-practice for future studies and developments.
7. Acknowledgements
The authors wish to thank Glenn Jocher for the
ideas of Mosaic data augmentation, the selection of
hyper-parameters by using genetic algorithms and solving
the grid sensitivity problem https://github.com/
ultralytics/yolov3.
10
Table 8: Comparison of the speed and accuracy of different object detectors on the MS COCO dataset (test-
dev 2017). (Real-time detectors with FPS 30 or higher are highlighted here. We compare the results with
batch=1 without using tensorRT.)
Method Backbone Size FPS AP AP50 AP75 APSAPMAPL
YOLOv4: Optimal Speed and Accuracy of Object Detection
YOLOv4 CSPDarknet-53 416 38 (M) 41.2% 62.8% 44.3% 20.4% 44.4% 56.0%
YOLOv4 CSPDarknet-53 512 31 (M) 43.0% 64.9% 46.5% 24.3% 46.1% 55.2%
YOLOv4 CSPDarknet-53 608 23 (M) 43.5% 65.7% 47.3% 26.7% 46.7% 53.3%
Learning Rich Features at High-Speed for Single-Shot Object Detection [84]
LRF VGG-16 300 76.9 (M) 32.0% 51.5% 33.8% 12.6% 34.9% 47.0%
LRF ResNet-101 300 52.6 (M) 34.3% 54.1% 36.6% 13.2% 38.2% 50.7%
LRF VGG-16 512 38.5 (M) 36.2% 56.6% 38.7% 19.0% 39.9% 48.8%
LRF ResNet-101 512 31.3 (M) 37.3% 58.5% 39.7% 19.7% 42.8% 50.1%
Receptive Field Block Net for Accurate and Fast Object Detection [47]
RFBNet VGG-16 300 66.7 (M) 30.3% 49.3% 31.8% 11.8% 31.9% 45.9%
RFBNet VGG-16 512 33.3 (M) 33.8% 54.2% 35.9% 16.2% 37.1% 47.4%
RFBNet-E VGG-16 512 30.3 (M) 34.4% 55.7% 36.4% 17.6% 37.0% 47.6%
YOLOv3: An incremental improvement [63]
YOLOv3 Darknet-53 320 45 (M) 28.2% 51.5% 29.7% 11.9% 30.6% 43.4%
YOLOv3 Darknet-53 416 35 (M) 31.0% 55.3% 32.3% 15.2% 33.2% 42.8%
YOLOv3 Darknet-53 608 20 (M) 33.0% 57.9% 34.4% 18.3% 35.4% 41.9%
YOLOv3-SPP Darknet-53 608 20 (M) 36.2% 60.6% 38.2% 20.6% 37.4% 46.1%
SSD: Single shot multibox detector [50]
SSD VGG-16 300 43 (M) 25.1% 43.1% 25.8% 6.6% 25.9% 41.4%
SSD VGG-16 512 22 (M) 28.8% 48.5% 30.3% 10.9% 31.8% 43.5%
Single-shot refinement neural network for object detection [95]
RefineDet VGG-16 320 38.7 (M) 29.4% 49.2% 31.3% 10.0% 32.0% 44.4%
RefineDet VGG-16 512 22.3 (M) 33.0% 54.5% 35.5% 16.3% 36.3% 44.3%
M2det: A single-shot object detector based on multi-level feature pyramid network [98]
M2det VGG-16 320 33.4 (M) 33.5% 52.4% 35.6% 14.4% 37.6% 47.6%
M2det ResNet-101 320 21.7 (M) 34.3% 53.5% 36.5% 14.8% 38.8% 47.9%
M2det VGG-16 512 18 (M) 37.6% 56.6% 40.5% 18.4% 43.4% 51.2%
M2det ResNet-101 512 15.8 (M) 38.8% 59.4% 41.7% 20.5% 43.9% 53.4%
M2det VGG-16 800 11.8 (M) 41.0% 59.7% 45.0% 22.1% 46.5% 53.8%
Parallel Feature Pyramid Network for Object Detection [34]
PFPNet-R VGG-16 320 33 (M) 31.8% 52.9% 33.6% 12% 35.5% 46.1%
PFPNet-R VGG-16 512 24 (M) 35.2% 57.6% 37.9% 18.7% 38.6% 45.9%
Focal Loss for Dense Object Detection [45]
RetinaNet ResNet-50 500 13.9 (M) 32.5% 50.9% 34.8% 13.9% 35.8% 46.7%
RetinaNet ResNet-101 500 11.1 (M) 34.4% 53.1% 36.8% 14.7% 38.5% 49.1%
RetinaNet ResNet-50 800 6.5 (M) 35.7% 55.0% 38.5% 18.9% 38.9% 46.3%
RetinaNet ResNet-101 800 5.1 (M) 37.8% 57.5% 40.8% 20.2% 41.1% 49.2%
Feature Selective Anchor-Free Module for Single-Shot Object Detection [102]
AB+FSAF ResNet-101 800 5.6 (M) 40.9% 61.5% 44.0% 24.0% 44.2% 51.3%
AB+FSAF ResNeXt-101 800 2.8 (M) 42.9% 63.8% 46.3% 26.6% 46.2% 52.7%
CornerNet: Detecting objects as paired keypoints [37]
CornerNet Hourglass 512 4.4 (M) 40.5% 57.8% 45.3% 20.8% 44.8% 56.7%
11
Table 9: Comparison of the speed and accuracy of different object detectors on the MS COCO dataset (test-dev 2017).
(Real-time detectors with FPS 30 or higher are highlighted here. We compare the results with batch=1 without using
tensorRT.)
Method Backbone Size FPS AP AP50 AP75 APSAPMAPL
YOLOv4: Optimal Speed and Accuracy of Object Detection
YOLOv4 CSPDarknet-53 416 54 (P) 41.2% 62.8% 44.3% 20.4% 44.4% 56.0%
YOLOv4 CSPDarknet-53 512 43 (P) 43.0% 64.9% 46.5% 24.3% 46.1% 55.2%
YOLOv4 CSPDarknet-53 608 33 (P) 43.5% 65.7% 47.3% 26.7% 46.7% 53.3%
CenterMask: Real-Time Anchor-Free Instance Segmentation [40]
CenterMask-Lite MobileNetV2-FPN 600×50.0 (P) 30.2% - - 14.2% 31.9% 40.9%
CenterMask-Lite VoVNet-19-FPN 600×43.5 (P) 35.9% - - 19.6% 38.0% 45.9%
CenterMask-Lite VoVNet-39-FPN 600×35.7 (P) 40.7% - - 22.4% 43.2% 53.5%
Enriched Feature Guided Refinement Network for Object Detection [57]
EFGRNet VGG-16 320 47.6 (P) 33.2% 53.4% 35.4% 13.4% 37.1% 47.9%
EFGRNet VG-G16 512 25.7 (P) 37.5% 58.8% 40.4% 19.7% 41.6% 49.4%
EFGRNet ResNet-101 512 21.7 (P) 39.0% 58.8% 42.3% 17.8% 43.6% 54.5%
Hierarchical Shot Detector [3]
HSD VGG-16 320 40 (P) 33.5% 53.2% 36.1% 15.0% 35.0% 47.8%
HSD VGG-16 512 23.3 (P) 38.8% 58.2% 42.5% 21.8% 41.9% 50.2%
HSD ResNet-101 512 20.8 (P) 40.2% 59.4% 44.0% 20.0% 44.4% 54.9%
HSD ResNeXt-101 512 15.2 (P) 41.9% 61.1% 46.2% 21.8% 46.6% 57.0%
HSD ResNet-101 768 10.9 (P) 42.3% 61.2% 46.9% 22.8% 47.3% 55.9%
Dynamic anchor feature selection for single-shot object detection [41]
DAFS VGG16 512 35 (P) 33.8% 52.9% 36.9% 14.6% 37.0% 47.7%
Soft Anchor-Point Object Detection [101]
SAPD ResNet-50 - 14.9 (P) 41.7% 61.9% 44.6% 24.1% 44.6% 51.6%
SAPD ResNet-50-DCN - 12.4 (P) 44.3% 64.4% 47.7% 25.5% 47.3% 57.0%
SAPD ResNet-101-DCN - 9.1 (P) 46.0% 65.9% 49.6% 26.3% 49.2% 59.6%
Region proposal by guided anchoring [82]
RetinaNet ResNet-50 - 10.8 (P) 37.1% 56.9% 40.0% 20.1% 40.1% 48.0%
Faster R-CNN ResNet-50 - 9.4 (P) 39.8% 59.2% 43.5% 21.8% 42.6% 50.7%
RepPoints: Point set representation for object detection [87]
RPDet ResNet-101 - 10 (P) 41.0% 62.9% 44.3% 23.6% 44.1% 51.7%
RPDet ResNet-101-DCN - 8 (P) 45.0% 66.1% 49.0% 26.6% 48.6% 57.5%
Libra R-CNN: Towards balanced learning for object detection [58]
Libra R-CNN ResNet-101 - 9.5 (P) 41.1% 62.1% 44.7% 23.4% 43.7% 52.5%
FreeAnchor: Learning to match anchors for visual object detection [96]
FreeAnchor ResNet-101 - 9.1 (P) 43.1% 62.2% 46.4% 24.5% 46.1% 54.8%
RetinaMask: Learning to Predict Masks Improves State-of-The-Art Single-Shot Detection for Free [14]
RetinaMask ResNet-50-FPN 800×8.1 (P) 39.4% 58.6% 42.3% 21.9% 42.0% 51.0%
RetinaMask ResNet-101-FPN 800×6.9 (P) 41.4% 60.8% 44.6% 23.0% 44.5% 53.5%
RetinaMask ResNet-101-FPN-GN 800×6.5 (P) 41.7% 61.7% 45.0% 23.5% 44.7% 52.8%
RetinaMask ResNeXt-101-FPN-GN 800×4.3 (P) 42.6% 62.5% 46.0% 24.8% 45.6% 53.8%
Cascade R-CNN: Delving into high quality object detection [2]
Cascade R-CNN ResNet-101 - 8 (P) 42.8% 62.1% 46.3% 23.7% 45.5% 55.2%
Centernet: Object detection with keypoint triplets [13]
Centernet Hourglass-52 - 4.4 (P) 41.6% 59.4% 44.2% 22.5% 43.1% 54.1%
Centernet Hourglass-104 - 3.3 (P) 44.9% 62.4% 48.1% 25.6% 47.4% 57.4%
Scale-Aware Trident Networks for Object Detection [42]
TridentNet ResNet-101 - 2.7 (P) 42.7% 63.6% 46.5% 23.9% 46.6% 56.6%
TridentNet ResNet-101-DCN - 1.3 (P) 46.8% 67.6% 51.5% 28.0% 51.2% 60.5%
12
Table 10: Comparison of the speed and accuracy of different object detectors on the MS COCO dataset (test-dev 2017).
(Real-time detectors with FPS 30 or higher are highlighted here. We compare the results with batch=1 without using
tensorRT.)
Method Backbone Size FPS AP AP50 AP75 APSAPMAPL
YOLOv4: Optimal Speed and Accuracy of Object Detection
YOLOv4 CSPDarknet-53 416 96 (V) 41.2% 62.8% 44.3% 20.4% 44.4% 56.0%
YOLOv4 CSPDarknet-53 512 83 (V) 43.0% 64.9% 46.5% 24.3% 46.1% 55.2%
YOLOv4 CSPDarknet-53 608 62 (V) 43.5% 65.7% 47.3% 26.7% 46.7% 53.3%
EfficientDet: Scalable and Efficient Object Detection [77]
EfficientDet-D0 Efficient-B0 512 62.5 (V) 33.8% 52.2% 35.8% 12.0% 38.3% 51.2%
EfficientDet-D1 Efficient-B1 640 50.0 (V) 39.6% 58.6% 42.3% 17.9% 44.3% 56.0%
EfficientDet-D2 Efficient-B2 768 41.7 (V) 43.0% 62.3% 46.2% 22.5% 47.0% 58.4%
EfficientDet-D3 Efficient-B3 896 23.8 (V) 45.8% 65.0% 49.3% 26.6% 49.4% 59.8%
Learning Spatial Fusion for Single-Shot Object Detection [48]
YOLOv3 + ASFF* Darknet-53 320 60 (V) 38.1% 57.4% 42.1% 16.1% 41.6% 53.6%
YOLOv3 + ASFF* Darknet-53 416 54 (V) 40.6% 60.6% 45.1% 20.3% 44.2% 54.1%
YOLOv3 + ASFF* Darknet-53 608×45.5 (V) 42.4% 63.0% 47.4% 25.5% 45.7% 52.3%
YOLOv3 + ASFF* Darknet-53 800×29.4 (V) 43.9% 64.1% 49.2% 27.0% 46.6% 53.4%
HarDNet: A Low Memory Traffic Network [4]
RFBNet HarDNet68 512 41.5 (V) 33.9% 54.3% 36.2% 14.7% 36.6% 50.5%
RFBNet HarDNet85 512 37.1 (V) 36.8% 57.1% 39.5% 16.9% 40.5% 52.9%
Focal Loss for Dense Object Detection [45]
RetinaNet ResNet-50 640 37 (V) 37.0% - - - - -
RetinaNet ResNet-101 640 29.4 (V) 37.9% - - - - -
RetinaNet ResNet-50 1024 19.6 (V) 40.1% - - - - -
RetinaNet ResNet-101 1024 15.4 (V) 41.1% - - - - -
SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection [88]
SM-NAS: E2 - 800×600 25.3 (V) 40.0% 58.2% 43.4% 21.1% 42.4% 51.7%
SM-NAS: E3 - 800×600 19.7 (V) 42.8% 61.2% 46.5% 23.5% 45.5% 55.6%
SM-NAS: E5 - 1333×800 9.3 (V) 45.9% 64.6% 49.6% 27.1% 49.0% 58.0%
NAS-FPN: Learning scalable feature pyramid architecture for object detection [17]
NAS-FPN ResNet-50 640 24.4 (V) 39.9% - - - - -
NAS-FPN ResNet-50 1024 12.7 (V) 44.2% - - - - -
Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection [94]
ATSS ResNet-101 800×17.5 (V) 43.6% 62.1% 47.4% 26.1% 47.0% 53.6%
ATSS ResNet-101-DCN 800×13.7 (V) 46.3% 64.7% 50.4% 27.7% 49.8% 58.4%
RDSNet: A New Deep Architecture for Reciprocal Object Detection and Instance Segmentation [83]
RDSNet ResNet-101 600 16.8 (V) 36.0% 55.2% 38.7% 17.4% 39.6% 49.7%
RDSNet ResNet-101 800 10.9 (V) 38.1% 58.5% 40.8% 21.2% 41.5% 48.2%
CenterMask: Real-Time Anchor-Free Instance Segmentation [40]
CenterMask ResNet-101-FPN 800×15.2 (V) 44.0% - - 25.8% 46.8% 54.9%
CenterMask VoVNet-99-FPN 800×12.9 (V) 46.5% - - 28.7% 48.9% 57.2%
13
References
[1] Navaneeth Bodla, Bharat Singh, Rama Chellappa, and
Larry S Davis. Soft-NMS–improving object detection with
one line of code. In Proceedings of the IEEE International
Conference on Computer Vision (ICCV), pages 5561–5569,
2017. 4
[2] Zhaowei Cai and Nuno Vasconcelos. Cascade R-CNN:
Delving into high quality object detection. In Proceedings
of the IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), pages 6154–6162, 2018. 12
[3] Jiale Cao, Yanwei Pang, Jungong Han, and Xuelong Li. Hi-
erarchical shot detector. In Proceedings of the IEEE In-
ternational Conference on Computer Vision (ICCV), pages
9705–9714, 2019. 12
[4] Ping Chao, Chao-Yang Kao, Yu-Shan Ruan, Chien-Hsiang
Huang, and Youn-Long Lin. HarDNet: A low memory traf-
fic network. Proceedings of the IEEE International Confer-
ence on Computer Vision (ICCV), 2019. 13
[5] Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos,
Kevin Murphy, and Alan L Yuille. DeepLab: Semantic im-
age segmentation with deep convolutional nets, atrous con-
volution, and fully connected CRFs. IEEE Transactions
on Pattern Analysis and Machine Intelligence (TPAMI),
40(4):834–848, 2017. 2,4
[6] Pengguang Chen. GridMask data augmentation. arXiv
preprint arXiv:2001.04086, 2020. 3
[7] Yukang Chen, Tong Yang, Xiangyu Zhang, Gaofeng Meng,
Xinyu Xiao, and Jian Sun. DetNAS: Backbone search for
object detection. In Advances in Neural Information Pro-
cessing Systems (NeurIPS), pages 6638–6648, 2019. 2
[8] Jiwoong Choi, Dayoung Chun, Hyun Kim, and Hyuk-Jae
Lee. Gaussian YOLOv3: An accurate and fast object de-
tector using localization uncertainty for autonomous driv-
ing. In Proceedings of the IEEE International Conference
on Computer Vision (ICCV), pages 502–511, 2019. 7
[9] Jifeng Dai, Yi Li, Kaiming He, and Jian Sun. R-FCN:
Object detection via region-based fully convolutional net-
works. In Advances in Neural Information Processing Sys-
tems (NIPS), pages 379–387, 2016. 2
[10] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li,
and Li Fei-Fei. ImageNet: A large-scale hierarchical im-
age database. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pages
248–255, 2009. 5
[11] Terrance DeVries and Graham W Taylor. Improved reg-
ularization of convolutional neural networks with CutOut.
arXiv preprint arXiv:1708.04552, 2017. 3
[12] Xianzhi Du, Tsung-Yi Lin, Pengchong Jin, Golnaz Ghiasi,
Mingxing Tan, Yin Cui, Quoc V Le, and Xiaodan Song.
SpineNet: Learning scale-permuted backbone for recog-
nition and localization. arXiv preprint arXiv:1912.05027,
2019. 2
[13] Kaiwen Duan, Song Bai, Lingxi Xie, Honggang Qi, Qing-
ming Huang, and Qi Tian. CenterNet: Keypoint triplets for
object detection. In Proceedings of the IEEE International
Conference on Computer Vision (ICCV), pages 6569–6578,
2019. 2,12
[14] Cheng-Yang Fu, Mykhailo Shvets, and Alexander C Berg.
RetinaMask: Learning to predict masks improves state-
of-the-art single-shot detection for free. arXiv preprint
arXiv:1901.03353, 2019. 12
[15] Robert Geirhos, Patricia Rubisch, Claudio Michaelis,
Matthias Bethge, Felix A Wichmann, and Wieland Brendel.
ImageNet-trained cnns are biased towards texture; increas-
ing shape bias improves accuracy and robustness. In Inter-
national Conference on Learning Representations (ICLR),
2019. 3
[16] Golnaz Ghiasi, Tsung-Yi Lin, and Quoc V Le. DropBlock:
A regularization method for convolutional networks. In Ad-
vances in Neural Information Processing Systems (NIPS),
pages 10727–10737, 2018. 3
[17] Golnaz Ghiasi, Tsung-Yi Lin, and Quoc V Le. NAS-FPN:
Learning scalable feature pyramid architecture for object
detection. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pages 7036–
7045, 2019. 2,13
[18] Ross Girshick. Fast R-CNN. In Proceedings of the IEEE In-
ternational Conference on Computer Vision (ICCV), pages
1440–1448, 2015. 2
[19] Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra
Malik. Rich feature hierarchies for accurate object de-
tection and semantic segmentation. In Proceedings of the
IEEE Conference on Computer Vision and Pattern Recog-
nition (CVPR), pages 580–587, 2014. 2,4
[20] Jianyuan Guo, Kai Han, Yunhe Wang, Chao Zhang, Zhao-
hui Yang, Han Wu, Xinghao Chen, and Chang Xu. Hit-
Detector: Hierarchical trinity architecture search for object
detection. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), 2020. 2
[21] Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing
Xu, and Chang Xu. GhostNet: More features from cheap
operations. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), 2020.
5
[22] Bharath Hariharan, Pablo Arbel´
aez, Ross Girshick, and
Jitendra Malik. Hypercolumns for object segmentation
and fine-grained localization. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), pages 447–456, 2015. 4
[23] Kaiming He, Georgia Gkioxari, Piotr Doll´
ar, and Ross Gir-
shick. Mask R-CNN. In Proceedings of the IEEE In-
ternational Conference on Computer Vision (ICCV), pages
2961–2969, 2017. 2
[24] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun.
Delving deep into rectifiers: Surpassing human-level per-
formance on ImageNet classification. In Proceedings of
the IEEE International Conference on Computer Vision
(ICCV), pages 1026–1034, 2015. 4
[25] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun.
Spatial pyramid pooling in deep convolutional networks for
visual recognition. IEEE Transactions on Pattern Analy-
sis and Machine Intelligence (TPAMI), 37(9):1904–1916,
2015. 2,4,7
[26] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun.
Deep residual learning for image recognition. In Proceed-
14
ings of the IEEE Conference on Computer Vision and Pat-
tern Recognition (CVPR), pages 770–778, 2016. 2
[27] Andrew Howard, Mark Sandler, Grace Chu, Liang-Chieh
Chen, Bo Chen, Mingxing Tan, Weijun Wang, Yukun Zhu,
Ruoming Pang, Vijay Vasudevan, et al. Searching for Mo-
bileNetV3. In Proceedings of the IEEE International Con-
ference on Computer Vision (ICCV), 2019. 2,4
[28] Andrew G Howard, Menglong Zhu, Bo Chen, Dmitry
Kalenichenko, Weijun Wang, Tobias Weyand, Marco An-
dreetto, and Hartwig Adam. MobileNets: Efficient con-
volutional neural networks for mobile vision applications.
arXiv preprint arXiv:1704.04861, 2017. 2,4
[29] Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation
networks. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pages 7132–
7141, 2018. 4
[30] Gao Huang, Zhuang Liu, Laurens Van Der Maaten, and Kil-
ian Q Weinberger. Densely connected convolutional net-
works. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pages 4700–
4708, 2017. 2
[31] Forrest N Iandola, Song Han, Matthew W Moskewicz,
Khalid Ashraf, William J Dally, and Kurt Keutzer.
SqueezeNet: AlexNet-level accuracy with 50x fewer pa-
rameters and¡ 0.5 MB model size. arXiv preprint
arXiv:1602.07360, 2016. 2
[32] Sergey Ioffe and Christian Szegedy. Batch normalization:
Accelerating deep network training by reducing internal co-
variate shift. arXiv preprint arXiv:1502.03167, 2015. 6
[33] Md Amirul Islam, Shujon Naha, Mrigank Rochan, Neil
Bruce, and Yang Wang. Label refinement network for
coarse-to-fine semantic segmentation. arXiv preprint
arXiv:1703.00551, 2017. 3
[34] Seung-Wook Kim, Hyong-Keun Kook, Jee-Young Sun,
Mun-Cheon Kang, and Sung-Jea Ko. Parallel feature pyra-
mid network for object detection. In Proceedings of the
European Conference on Computer Vision (ECCV), pages
234–250, 2018. 11
[35] G¨
unter Klambauer, Thomas Unterthiner, Andreas Mayr,
and Sepp Hochreiter. Self-normalizing neural networks.
In Advances in Neural Information Processing Systems
(NIPS), pages 971–980, 2017. 4
[36] Gustav Larsson, Michael Maire, and Gregory
Shakhnarovich. FractalNet: Ultra-deep neural net-
works without residuals. arXiv preprint arXiv:1605.07648,
2016. 6
[37] Hei Law and Jia Deng. CornerNet: Detecting objects as
paired keypoints. In Proceedings of the European Confer-
ence on Computer Vision (ECCV), pages 734–750, 2018. 2,
11
[38] Hei Law, Yun Teng, Olga Russakovsky, and Jia Deng.
CornerNet-Lite: Efficient keypoint based object detection.
arXiv preprint arXiv:1904.08900, 2019. 2
[39] Svetlana Lazebnik, Cordelia Schmid, and Jean Ponce. Be-
yond bags of features: Spatial pyramid matching for recog-
nizing natural scene categories. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), volume 2, pages 2169–2178. IEEE, 2006. 4
[40] Youngwan Lee and Jongyoul Park. CenterMask: Real-time
anchor-free instance segmentation. In Proceedings of the
IEEE Conference on Computer Vision and Pattern Recog-
nition (CVPR), 2020. 12,13
[41] Shuai Li, Lingxiao Yang, Jianqiang Huang, Xian-Sheng
Hua, and Lei Zhang. Dynamic anchor feature selection for
single-shot object detection. In Proceedings of the IEEE In-
ternational Conference on Computer Vision (ICCV), pages
6609–6618, 2019. 12
[42] Yanghao Li, Yuntao Chen, Naiyan Wang, and Zhaoxiang
Zhang. Scale-aware trident networks for object detection.
In Proceedings of the IEEE International Conference on
Computer Vision (ICCV), pages 6054–6063, 2019. 12
[43] Zeming Li, Chao Peng, Gang Yu, Xiangyu Zhang, Yang-
dong Deng, and Jian Sun. DetNet: Design backbone for
object detection. In Proceedings of the European Confer-
ence on Computer Vision (ECCV), pages 334–350, 2018.
2
[44] Tsung-Yi Lin, Piotr Doll´
ar, Ross Girshick, Kaiming He,
Bharath Hariharan, and Serge Belongie. Feature pyramid
networks for object detection. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), pages 2117–2125, 2017. 2
[45] Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He,
and Piotr Doll´
ar. Focal loss for dense object detection. In
Proceedings of the IEEE International Conference on Com-
puter Vision (ICCV), pages 2980–2988, 2017. 2,3,11,13
[46] Tsung-Yi Lin, Michael Maire, Serge Belongie, James
Hays, Pietro Perona, Deva Ramanan, Piotr Doll´
ar, and
C Lawrence Zitnick. Microsoft COCO: Common objects
in context. In Proceedings of the European Conference on
Computer Vision (ECCV), pages 740–755, 2014. 5
[47] Songtao Liu, Di Huang, et al. Receptive field block net for
accurate and fast object detection. In Proceedings of the
European Conference on Computer Vision (ECCV), pages
385–400, 2018. 2,4,11
[48] Songtao Liu, Di Huang, and Yunhong Wang. Learning spa-
tial fusion for single-shot object detection. arXiv preprint
arXiv:1911.09516, 2019. 2,4,13
[49] Shu Liu, Lu Qi, Haifang Qin, Jianping Shi, and Jiaya Jia.
Path aggregation network for instance segmentation. In
Proceedings of the IEEE Conference on Computer Vision
and Pattern Recognition (CVPR), pages 8759–8768, 2018.
1,2,7
[50] Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian
Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C
Berg. SSD: Single shot multibox detector. In Proceedings
of the European Conference on Computer Vision (ECCV),
pages 21–37, 2016. 2,11
[51] Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully
convolutional networks for semantic segmentation. In Pro-
ceedings of the IEEE Conference on Computer Vision and
Pattern Recognition (CVPR), pages 3431–3440, 2015. 4
[52] Ilya Loshchilov and Frank Hutter. SGDR: Stochas-
tic gradient descent with warm restarts. arXiv preprint
arXiv:1608.03983, 2016. 7
[53] Ningning Ma, Xiangyu Zhang, Hai-Tao Zheng, and Jian
Sun. ShuffleNetV2: Practical guidelines for efficient cnn
15
architecture design. In Proceedings of the European Con-
ference on Computer Vision (ECCV), pages 116–131, 2018.
2
[54] Andrew L Maas, Awni Y Hannun, and Andrew Y Ng. Rec-
tifier nonlinearities improve neural network acoustic mod-
els. In Proceedings of International Conference on Ma-
chine Learning (ICML), volume 30, page 3, 2013. 4
[55] Diganta Misra. Mish: A self regularized non-
monotonic neural activation function. arXiv preprint
arXiv:1908.08681, 2019. 4
[56] Vinod Nair and Geoffrey E Hinton. Rectified linear units
improve restricted boltzmann machines. In Proceedings
of International Conference on Machine Learning (ICML),
pages 807–814, 2010. 4
[57] Jing Nie, Rao Muhammad Anwer, Hisham Cholakkal, Fa-
had Shahbaz Khan, Yanwei Pang, and Ling Shao. Enriched
feature guided refinement network for object detection. In
Proceedings of the IEEE International Conference on Com-
puter Vision (ICCV), pages 9537–9546, 2019. 12
[58] Jiangmiao Pang, Kai Chen, Jianping Shi, Huajun Feng,
Wanli Ouyang, and Dahua Lin. Libra R-CNN: Towards bal-
anced learning for object detection. In Proceedings of the
IEEE Conference on Computer Vision and Pattern Recog-
nition (CVPR), pages 821–830, 2019. 2,12
[59] Prajit Ramachandran, Barret Zoph, and Quoc V Le.
Searching for activation functions. arXiv preprint
arXiv:1710.05941, 2017. 4
[60] Abdullah Rashwan, Agastya Kalra, and Pascal Poupart.
Matrix Nets: A new deep architecture for object detection.
In Proceedings of the IEEE International Conference on
Computer Vision Workshop (ICCV Workshop), pages 0–0,
2019. 2
[61] Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali
Farhadi. You only look once: Unified, real-time object de-
tection. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pages 779–
788, 2016. 2
[62] Joseph Redmon and Ali Farhadi. YOLO9000: better, faster,
stronger. In Proceedings of the IEEE Conference on Com-
puter Vision and Pattern Recognition (CVPR), pages 7263–
7271, 2017. 2
[63] Joseph Redmon and Ali Farhadi. YOLOv3: An incremental
improvement. arXiv preprint arXiv:1804.02767, 2018. 2,
4,7,11
[64] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun.
Faster R-CNN: Towards real-time object detection with re-
gion proposal networks. In Advances in Neural Information
Processing Systems (NIPS), pages 91–99, 2015. 2
[65] Hamid Rezatofighi, Nathan Tsoi, JunYoung Gwak, Amir
Sadeghian, Ian Reid, and Silvio Savarese. Generalized in-
tersection over union: A metric and a loss for bounding
box regression. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pages
658–666, 2019. 3
[66] Mark Sandler, Andrew Howard, Menglong Zhu, Andrey
Zhmoginov, and Liang-Chieh Chen. MobileNetV2: In-
verted residuals and linear bottlenecks. In Proceedings
of the IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), pages 4510–4520, 2018. 2
[67] Abhinav Shrivastava, Abhinav Gupta, and Ross Girshick.
Training region-based object detectors with online hard ex-
ample mining. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pages
761–769, 2016. 3
[68] Karen Simonyan and Andrew Zisserman. Very deep convo-
lutional networks for large-scale image recognition. arXiv
preprint arXiv:1409.1556, 2014. 2
[69] Krishna Kumar Singh, Hao Yu, Aron Sarmasi, Gautam
Pradeep, and Yong Jae Lee. Hide-and-Seek: A data aug-
mentation technique for weakly-supervised localization and
beyond. arXiv preprint arXiv:1811.02545, 2018. 3
[70] Saurabh Singh and Shankar Krishnan. Filter response
normalization layer: Eliminating batch dependence in
the training of deep neural networks. arXiv preprint
arXiv:1911.09737, 2019. 6
[71] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya
Sutskever, and Ruslan Salakhutdinov. DropOut: A simple
way to prevent neural networks from overfitting. The jour-
nal of machine learning research, 15(1):1929–1958, 2014.
3
[72] K-K Sung and Tomaso Poggio. Example-based learning
for view-based human face detection. IEEE Transactions
on Pattern Analysis and Machine Intelligence (TPAMI),
20(1):39–51, 1998. 3
[73] Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jon
Shlens, and Zbigniew Wojna. Rethinking the inception ar-
chitecture for computer vision. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), pages 2818–2826, 2016. 3
[74] Mingxing Tan, Bo Chen, Ruoming Pang, Vijay Vasudevan,
Mark Sandler, Andrew Howard, and Quoc V Le. MNAS-
net: Platform-aware neural architecture search for mobile.
In Proceedings of the IEEE Conference on Computer Vision
and Pattern Recognition (CVPR), pages 2820–2828, 2019.
2
[75] Mingxing Tan and Quoc V Le. EfficientNet: Rethinking
model scaling for convolutional neural networks. In Pro-
ceedings of International Conference on Machine Learning
(ICML), 2019. 2
[76] Mingxing Tan and Quoc V Le. MixNet: Mixed depthwise
convolutional kernels. In Proceedings of the British Ma-
chine Vision Conference (BMVC), 2019. 5
[77] Mingxing Tan, Ruoming Pang, and Quoc V Le. Efficient-
Det: Scalable and efficient object detection. In Proceedings
of the IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), 2020. 2,4,13
[78] Zhi Tian, Chunhua Shen, Hao Chen, and Tong He. FCOS:
Fully convolutional one-stage object detection. In Proceed-
ings of the IEEE International Conference on Computer Vi-
sion (ICCV), pages 9627–9636, 2019. 2
[79] Jonathan Tompson, Ross Goroshin, Arjun Jain, Yann Le-
Cun, and Christoph Bregler. Efficient object localization
using convolutional networks. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), pages 648–656, 2015. 6
16
[80] Li Wan, Matthew Zeiler, Sixin Zhang, Yann Le Cun, and
Rob Fergus. Regularization of neural networks using Drop-
Connect. In Proceedings of International Conference on
Machine Learning (ICML), pages 1058–1066, 2013. 3
[81] Chien-Yao Wang, Hong-Yuan Mark Liao, Yueh-Hua Wu,
Ping-Yang Chen, Jun-Wei Hsieh, and I-Hau Yeh. CSPNet:
A new backbone that can enhance learning capability of
cnn. Proceedings of the IEEE Conference on Computer Vi-
sion and Pattern Recognition Workshop (CVPR Workshop),
2020. 2,7
[82] Jiaqi Wang, Kai Chen, Shuo Yang, Chen Change Loy, and
Dahua Lin. Region proposal by guided anchoring. In Pro-
ceedings of the IEEE Conference on Computer Vision and
Pattern Recognition (CVPR), pages 2965–2974, 2019. 12
[83] Shaoru Wang, Yongchao Gong, Junliang Xing, Lichao
Huang, Chang Huang, and Weiming Hu. RDSNet: A
new deep architecture for reciprocal object detection and
instance segmentation. arXiv preprint arXiv:1912.05070,
2019. 13
[84] Tiancai Wang, Rao Muhammad Anwer, Hisham Cholakkal,
Fahad Shahbaz Khan, Yanwei Pang, and Ling Shao. Learn-
ing rich features at high-speed for single-shot object detec-
tion. In Proceedings of the IEEE International Conference
on Computer Vision (ICCV), pages 1971–1980, 2019. 11
[85] Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In
So Kweon. CBAM: Convolutional block attention module.
In Proceedings of the European Conference on Computer
Vision (ECCV), pages 3–19, 2018. 1,2,4
[86] Saining Xie, Ross Girshick, Piotr Doll´
ar, Zhuowen Tu, and
Kaiming He. Aggregated residual transformations for deep
neural networks. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pages
1492–1500, 2017. 2
[87] Ze Yang, Shaohui Liu, Han Hu, Liwei Wang, and Stephen
Lin. RepPoints: Point set representation for object detec-
tion. In Proceedings of the IEEE International Conference
on Computer Vision (ICCV), pages 9657–9666, 2019. 2,12
[88] Lewei Yao, Hang Xu, Wei Zhang, Xiaodan Liang, and
Zhenguo Li. SM-NAS: Structural-to-modular neural archi-
tecture search for object detection. In Proceedings of the
AAAI Conference on Artificial Intelligence (AAAI), 2020.
13
[89] Zhuliang Yao, Yue Cao, Shuxin Zheng, Gao Huang, and
Stephen Lin. Cross-iteration batch normalization. arXiv
preprint arXiv:2002.05712, 2020. 1,6
[90] Jiahui Yu, Yuning Jiang, Zhangyang Wang, Zhimin Cao,
and Thomas Huang. UnitBox: An advanced object detec-
tion network. In Proceedings of the 24th ACM international
conference on Multimedia, pages 516–520, 2016. 3
[91] Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk
Chun, Junsuk Choe, and Youngjoon Yoo. CutMix: Regu-
larization strategy to train strong classifiers with localizable
features. In Proceedings of the IEEE International Confer-
ence on Computer Vision (ICCV), pages 6023–6032, 2019.
3
[92] Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and
David Lopez-Paz. MixUp: Beyond empirical risk mini-
mization. arXiv preprint arXiv:1710.09412, 2017. 3
[93] Hang Zhang, Kristin Dana, Jianping Shi, Zhongyue Zhang,
Xiaogang Wang, Ambrish Tyagi, and Amit Agrawal. Con-
text encoding for semantic segmentation. In Proceedings
of the IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), pages 7151–7160, 2018. 6
[94] Shifeng Zhang, Cheng Chi, Yongqiang Yao, Zhen Lei, and
Stan Z Li. Bridging the gap between anchor-based and
anchor-free detection via adaptive training sample selec-
tion. In Proceedings of the IEEE Conference on Computer
Vision and Pattern Recognition (CVPR), 2020. 13
[95] Shifeng Zhang, Longyin Wen, Xiao Bian, Zhen Lei, and
Stan Z Li. Single-shot refinement neural network for ob-
ject detection. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pages
4203–4212, 2018. 11
[96] Xiaosong Zhang, Fang Wan, Chang Liu, Rongrong Ji, and
Qixiang Ye. FreeAnchor: Learning to match anchors for
visual object detection. In Advances in Neural Information
Processing Systems (NeurIPS), 2019. 12
[97] Xiangyu Zhang, Xinyu Zhou, Mengxiao Lin, and Jian Sun.
ShuffleNet: An extremely efficient convolutional neural
network for mobile devices. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), pages 6848–6856, 2018. 2
[98] Qijie Zhao, Tao Sheng, Yongtao Wang, Zhi Tang, Ying
Chen, Ling Cai, and Haibin Ling. M2det: A single-shot
object detector based on multi-level feature pyramid net-
work. In Proceedings of the AAAI Conference on Artificial
Intelligence (AAAI), volume 33, pages 9259–9266, 2019. 2,
4,11
[99] Zhaohui Zheng, Ping Wang, Wei Liu, Jinze Li, Rongguang
Ye, and Dongwei Ren. Distance-IoU Loss: Faster and bet-
ter learning for bounding box regression. In Proceedings
of the AAAI Conference on Artificial Intelligence (AAAI),
2020. 3,4
[100] Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li,
and Yi Yang. Random erasing data augmentation. arXiv
preprint arXiv:1708.04896, 2017. 3
[101] Chenchen Zhu, Fangyi Chen, Zhiqiang Shen, and Mar-
ios Savvides. Soft anchor-point object detection. arXiv
preprint arXiv:1911.12448, 2019. 12
[102] Chenchen Zhu, Yihui He, and Marios Savvides. Feature se-
lective anchor-free module for single-shot object detection.
In Proceedings of the IEEE Conference on Computer Vision
and Pattern Recognition (CVPR), pages 840–849, 2019. 11
17
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
Bounding box regression is the crucial step in object detection. In existing methods, while ℓn-norm loss is widely adopted for bounding box regression, it is not tailored to the evaluation metric, i.e., Intersection over Union (IoU). Recently, IoU loss and generalized IoU (GIoU) loss have been proposed to benefit the IoU metric, but still suffer from the problems of slow convergence and inaccurate regression. In this paper, we propose a Distance-IoU (DIoU) loss by incorporating the normalized distance between the predicted box and the target box, which converges much faster in training than IoU and GIoU losses. Furthermore, this paper summarizes three geometric factors in bounding box regression, i.e., overlap area, central point distance and aspect ratio, based on which a Complete IoU (CIoU) loss is proposed, thereby leading to faster convergence and better performance. By incorporating DIoU and CIoU losses into state-of-the-art object detection algorithms, e.g., YOLO v3, SSD and Faster R-CNN, we achieve notable performance gains in terms of not only IoU metric but also GIoU metric. Moreover, DIoU can be easily adopted into non-maximum suppression (NMS) to act as the criterion, further boosting performance improvement. The source code and trained models are available at https://github.com/Zzh-tju/DIoU
Article
Object detection and instance segmentation are two fundamental computer vision tasks. They are closely correlated but their relations have not yet been fully explored in most previous work. This paper presents RDSNet, a novel deep architecture for reciprocal object detection and instance segmentation. To reciprocate these two tasks, we design a two-stream structure to learn features on both the object level (i.e., bounding boxes) and the pixel level (i.e., instance masks) jointly. Within this structure, information from the two streams is fused alternately, namely information on the object level introduces the awareness of instance and translation variance to the pixel level, and information on the pixel level refines the localization accuracy of objects on the object level in return. Specifically, a correlation module and a cropping module are proposed to yield instance masks, as well as a mask based boundary refinement module for more accurate bounding boxes. Extensive experimental analyses and comparisons on the COCO dataset demonstrate the effectiveness and efficiency of RDSNet. The source code is available at https://github.com/wangsr126/RDSNet.