Navneet Dalal's scientific contributions

Publications (11)

Conference Paper
Detecting humans in films and videos is a challenging problem owing to the motion of the subjects, the camera and the background and to varia- tions in pose, appearance, clothing, illumination and background clutter. We de- velop a detector for standing and moving people in videos with possibly moving cameras and backgrounds, testing several differ...
Article
Full-text available
View interpolation has been explored in the scientific community as a means to avoid the complexity of full 3D in the construction of photo-realistic interactive scenarios. EVENTS project attempts to apply state of the art view interpolation to the field of professional sports. The aim is to populate a wide scenario such as a stadium with a number...
Conference Paper
We study the question of feature sets for robust visual object recognition, adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of Histograms of Oriented Gradient (HOG) descriptors significantly outperform existing feature sets for human detection....
Conference Paper
Full-text available
The PASCAL Visual Object Classes Challenge ran from February to March 2005. The goal of the challenge was to recognize objects from a number of visual object classes in realistic scenes (i.e. not pre-segmented objects). Four object classes were selected: motor- bikes, bicycles, cars and people. Twelve teams entered the challenge. In this chapter we...
Article
In this paper we describe a method for analysing video sequences and for representing them as mosaics or panoramas. Previous work on video mosaicking essentially concentrated on static scenes. We generalize these approaches to the case of a rotating camera observing both static and moving objects where the static portions of the scene are not neces...
Conference Paper
This paper proposes a robust estimation and validation framework for characterizing local structures in a positive multi-variate continuous function approximated by a Gaussian-based model. The new solution is robust against data with large deviations from the model and margin-truncations induced by neighboring structures. To this goal, it unifies r...
Conference Paper
Full-text available
Given two or more video sequences containing similar human activities (running, jumping, etc.), we want to devise a method which extracts spatio-temporal signatures associated with these activities, compares these signatures, and aligns key positions from different videos. We introduce a method which, in conjunction with a number of hypotheses, all...
Conference Paper
Full-text available
We address the problem of constructing mosaics from video sequences taken by rotating cameras. In particular we investigate the widespread case where the scene is not only static but may also contain large dynamic areas, induced by moving or deforming objects. Most of the existing techniques fail to produce reliable results on such video sequences....
Article
Given two or more video sequences containing similar human activities (running, jumping, etc.) we want to devise a method which extracts spatio-temporal signatures associated with these activities, compares these signatures, and aligns key positions of different videos. In this paper we introduce a method which, in conjunction with a number of hypo...

Citations

... Dalal & Triggs, 2005) pretrained on a large set of faces under highly variable conditions of expression and environmental factors (300-W database;Sagonas et al., 2016). ...
... FF++ dataset provides a benchmark with 1000 hidden images manipulated by these four methods. 7 In this work, we applied Scale Invariant Feature Transform(SIFT) [29] and Histograms of Oriented Gradients(HOG) [12] to extract handcrafted features. Then we feed these features to a support vector machine (SVM) for binary classification. ...
... In graphics, image mosaics play an important role in the field of image based rendering, which aims to render photorealistic views from collections of real world images [3], [4]. For applications such as virtual travel and architectural walkthroughs, it is desirable to have complete panoramas, i.e., mosaics which cover the whole viewing sphere and hence allow the user to look in any direction [10]. Image alignment algorithms can discover the correspondence relationships among images with varying degrees of overlap. ...
... With the current technology, it is not possible to capture images with a large field of view and a good resolution, that is why mosaicking methods are used to get a better mosaic quality. In general, there are two types of mosaics, the first one uses static methods where the scene doesn't include any motion [1,2,3,4,5] while the latter uses dynamic methods where the scene contains different dynamic events [6,7,8,9]. Furthermore, image mosaics (panoramas) are widely used in many fields of computer vision where it is necessary to have a larger field ...
... In the ablation experiments, to evaluate the classification and localization performance and computational complexity of our designed modules, in terms of the definitions of VOC2007 and COCO metrics [40,41], we make the metrics as shown in Table 1: ...
... A comparison is done between the manifestation of the CT nodules and the standard Gaussian intensity model. Okato [27,28] used a Robust model for fitting anisotropic Gaussian model and Mean Shift algorithm for fitting at each scale. In Diciotti et al [29] the nodules are segmented using multiscale LoG filtering. ...
... To deal with these challenges, most previous methods focus on learning how to describe normal events by training models using only normal events, and then the events that get bad descriptions are considered as anomalous events. Early works usually utilize handcrafted features to describe events in video sequences, such as histogram of oriented gradients (HOG) [5], histogram of optical flow (HOF) [6] and mixture of dynamic textures (MDTs) [17]. Nevertheless, the aforementioned handcrafted features are extracted by deterministic filters and are not able to adapt to different complex scenarios. ...
... A few of them [27][28][29] face the problems of automatic camera calibration and player recognition for the estimation of distances in the playing field, the provision of overlaid data on the game statistics, the forwarding of live event alerts. Some works propose methods for providing users with enhanced images of the playing field such as photo realistic views, virtual flights through real soccer scenes, and mosaic images [30][31][32]. ...
... GMC is an essential module for processing videos from non-stationary cameras, which are abundant due to emerging mobile sensors, e.g., wearable cameras, smartphones, and camera drones. First, the resultant motion panorama [3], as if virtually generated by a static camera, is by itself appealing for visual perception. More importantly, many vision tasks may benefit from GMC. ...
... As a concrete example, we consider high jump. We can automatically collect the athlete's center of mass information from video and convert the data into a time series (It is possible to correct for the cameras pan and tilt; see [8]). We found that when we issued queries to a database of high jumps, we got intuitive answers only when doing SWM. ...