Computer Vision

Follow this topic

Question:
Open Where can I get the dataset common color MPEG-7 used in the core experiments?
I need images for my tests.

I need images for my tests.

By Carlos Pérez Lara · Centro Nacional de Investigación y Desarrollo Tecnológico

Following
Answer added to:
4 Is there a good survey on the function of V1, V2 and V4 of the human cortex?
By Robert Munnoch · The University of Hull

Following
Question:
Open GASAC (Genetic Algorithm Sample Consensus)
Does anyone can help me to find working code for GASAC (Genetic Algorithm Sample Consensus) algorithm? Matlab implementation is also welcome.

Does anyone can help me to find working code for GASAC (Genetic Algorithm Sample Consensus) algorithm? Matlab implementation is also welcome.

By Iva Nikolova · Technical University of Sofia

Following
Answer added to:
12 What is the best unsupervised method for feature subset selection?
By Devanjali Relan · The University of Edinburgh

Abdul Momin · Indian Institute of Information Technology Allahabad

"Best method ?" .... I don't think any of the algorithm is best. These algorithms works best for some particular condition.So first analysis your data... [more]

"Best method ?" .... I don't think any of the algorithm is best. These algorithms works best for some particular condition.So first analysis your data and analyze which of the algorithm is working best for them. "None is perfect"

Following
Answer added to:
13 Is there any tutorial for implementing image processing algorithms over GPU?
By Ali Awad · Kyushu University

Sajid Farooq · University of Glasgow

Which language are you using? Depending on the language, there are a multitude of tutorials out there, for GLSL, CUDA, as well as OpenCL.

Which language are you using? Depending on the language, there are a multitude of tutorials out there, for GLSL, CUDA, as well as OpenCL.

Following
Answer added to:
5 Can anyone tell me about an algorithm/methodology to match the feature points between two images which is invariant to scale, rotation and illumination in opencv2.1?
By Rashmi Chinnappa · Sri Jayachamarajendra College of Engineering

Mahesh Chikkanna · National Institute of Technology Karnataka

SIFT alg can be used check it out

SIFT alg can be used check it out

Following
Answer added to:
4 Can anyone suggest a measure for range images registration?
By Mahdi Manafzade · Ferdowsi University Of Mashhad

Daniel Schwarz · Masaryk University

Consider combining the measures based on confusion matrix with true positives, false positives, true negatives and false negatives. This gives you pos... [more]

Consider combining the measures based on confusion matrix with true positives, false positives, true negatives and false negatives. This gives you possibility to account also for cases when a match is done but not the desired one etc... Look for terms confusion matrix, specificity, sensitivity.

Following
Answer added to:
3 Matching descriptors: other choices
By Néstor Morales Hernández · Universidad de La Laguna

Pierre Moulon · Université Paris-Est Marne-la-Vallée

You can take a look to quantization based approach: - Vocabulary tree, - Product quantization for nearest neighbor search, - LSH, Multi Probe LSH meth... [more]

You can take a look to quantization based approach: - Vocabulary tree, - Product quantization for nearest neighbor search, - LSH, Multi Probe LSH method, - Domain transform (http://cvlab.epfl.ch/research/detect/ldahash)

Following
Answer added to:
5 In order to perform registration of two images which only differ slightly at the viewpoint is the affine transformation adequate?
By Evangelos Skodras · University of Patras

Christos Stentoumis · National Technical University of Athens

I take as a fact that your views are not from the same position. You could have same position if you only rotated your cameras or your subject moved b... [more]

I take as a fact that your views are not from the same position. You could have same position if you only rotated your cameras or your subject moved but not your camera. So again if fig.3 is valid, then you could use a pipeline where keypoints in images are found and matched. Then the relative orientation of the stereo pairs could be retrieved. There are structure from motion algorithms open source distributed to help you. If you have the luxury of pre-calibrating the cameras, more/different things can be done. For example radial distortion could be found from a check-board pattern to use it in your solution. If I am on the right track for your subject, we could discuss more details for a more standard pipeline.

Following
Answer added to:
6 In OpenCV, which are the fastest 2D feature detector and extractor algorithms?
By Massimo Mauro · Università degli Studi di Brescia

Robert Munnoch · The University of Hull

If one is only looking for feature location I have found FAST to be useful and very fast... No pun intended it is also very discriminating with well ... [more]

If one is only looking for feature location I have found FAST to be useful and very fast... No pun intended it is also very discriminating with well chosen thresholds 1-2 ms seen on an android device for a 640*480 image even faster on pc

Following
Answer added to:
4 What are the best algorithms for non-rigid point set registration?
By Néstor Morales Hernández · Universidad de La Laguna

Néstor Morales Hernández · Universidad de La Laguna

Thank you. However, I am thinking in ICP-like algorithms for non-rigid registration that doesn't require the use of descriptors for the alignment of t... [more]

Thank you. However, I am thinking in ICP-like algorithms for non-rigid registration that doesn't require the use of descriptors for the alignment of the point clouds.

Following
Answer added to:
3 Is anyone working on pedestrian group detection and the effect of past data on 'real time group detection and scene interpretation'?
By Arun Chandran · National University of Singapore

Martin Lukac · Tohoku University

I do not know about any such research or approaches as to me personally it seems a bit stretched. The reason is the following. If I have a set of obje... [more]

I do not know about any such research or approaches as to me personally it seems a bit stretched. The reason is the following. If I have a set of object with nearly parallel trajectories they become implicitly a group. What I do not understand really is why you are detecting it as a group of pedestrians instead of just detecting a set of moving objects, identify them as people and then you have pedestrians.

Following
Answer added to:
4 What are the best stereo matching algorithms for real time stereo vision?
By Youssef Fathi · Université Moulay Ismail

Dibya Mohanty · Indian Statistical Institute

I have used DAISY descriptor for the same purpose in my work. You can go here ( http://cvlab.epfl.ch/software/daisy ) and check it's helpful

I have used DAISY descriptor for the same purpose in my work. You can go here ( http://cvlab.epfl.ch/software/daisy ) and check it's helpful

Following
Answer added to:
10 Efficient way for image alignment by MATLAB??
By Humayra Ali · Flinders University

Mahdi Manafzade · Ferdowsi University Of Mashhad

In my opinion, feature based methods like SIFT and SURF are good enough for alignment. The methods has been implemented in MATLAB.

In my opinion, feature based methods like SIFT and SURF are good enough for alignment. The methods has been implemented in MATLAB.

Following
Answer added to:
4 I'm interested in finding information regarding binocular disparity.
By Katie Stern · University of Wisconsin - Stevens Point

Katie Stern · University of Wisconsin - Stevens Point

Thies, I am so grateful for your response! Yes, please post papers whenever you have the chance. Thank you! Katie

Thies, I am so grateful for your response! Yes, please post papers whenever you have the chance. Thank you! Katie

Following
Question:
Open Feature function for conditional random field in sequence detection of gesture recognition.
How do you use the feature function in for gesture recognition? Since the values recorded will change during the training.

How do you use the feature function in for gesture recognition? Since the values recorded will change during the training.

By Mahesh Chikkanna · National Institute of Technology Karnataka

Following
Answer added to:
2 How to reconstruct a 3D image from two 2D images of the same scene taken from the same camera?
By Rachakonda Poojitha

Pierre Moulon · Université Paris-Est Marne-la-Vallée

Once you have a P = [K [R|t]] for each camera for each correspondence you have to triangulate in order to have a 3D point. => i.e attachment You will... [more]

Once you have a P = [K [R|t]] for each camera for each correspondence you have to triangulate in order to have a 3D point. => i.e attachment You will find code and examples here: http://imagine.enpc.fr/~moulonp/openMVG/ I think to this example (Essential Matrix + triangulation) https://github.com/openMVG/openMVG/tree/master/src/openMVG_Samples/robust_essential In order to have good quality 3D points your camera must have a not null baseline.

http://imagine.enpc.fr/~moulonp/openMVG/Triangulation_n_geometry.png ×

Following
Answer added to:
5 What are the most common edge based segmentation methods?
By Michal Lisicki · Warsaw University of Technology

Amine Larhmam · Université de Mons

For some basicson image processing you can start using imagej which have different plugin. Also, there is an online course specialized on image proces... [more]

For some basicson image processing you can start using imagej which have different plugin. Also, there is an online course specialized on image processing in coursera.org. I found it very interesting : https://class.coursera.org/images-2012-001/class/index Good luck

Following
Answer added to:
2 Is there a good paper or book about tensor application in computer vision?
By Xiaohua Huang · University of Oulu

Xiaohua Huang · University of Oulu

Thanks Daniel. This is a good material for me.

Thanks Daniel. This is a good material for me.

Following
Question:
Open Anyone working or worked with functional difference predictors (FDPs)?
Does anyone know some follow-up work on the original paper by James A. Ferwerda?

Does anyone know some follow-up work on the original paper by James A. Ferwerda?

By Stephan Irgenfried · Karlsruhe Institute of Technology

Following
Question:
Open Bird's eye view Map
Can anyone help me with the code for constructing a bird's eye view for robotic mapping?

Can anyone help me with the code for constructing a bird's eye view for robotic mapping?

By Yunusa Sai'd · Putra University, Malaysia

Following
Answer added to:
3 Data Analysis Tools for Researchers (especially Computer Vision Engineering)
By Anh Cat Le Ngo · University of Nottingham, Malaysia Campus

Anh Cat Le Ngo · University of Nottingham, Malaysia Campus

@Ankur Kumar: I just know about OpenCV and totally blind about VTK, ITK and PCL. Do you mind enlighten me on those tools such as their purposes and us... [more]

@Ankur Kumar: I just know about OpenCV and totally blind about VTK, ITK and PCL. Do you mind enlighten me on those tools such as their purposes and usage case etc ?

Following
Question:
Open How to "increase PSNR of super resolved" image using watermarking"
I want to increase PSNR of SR image by embedding watermark in it.

I want to increase PSNR of SR image by embedding watermark in it.

By Vaibhav Joshi · Dhirubhai Ambani Institute of Information and Communication Technology

Following
Answer added to:
2 What is the best corner detection data-set ?
By seyed mostafa mousavi kahaki · National University of Malaysia

seyed mostafa mousavi kahaki · National University of Malaysia

Thanks dinesh for your answer, yes these images are good and from standard dataset, but i need the complete dataset with different image quality. Btw ... [more]

Thanks dinesh for your answer, yes these images are good and from standard dataset, but i need the complete dataset with different image quality. Btw thanks.

Following
Answer added to:
4 Using an android phone camera as an imaging device for a laptop.
By Tousif Zaman · CMR Institute of Technology

Tousif Zaman · CMR Institute of Technology

Thank you for all the suggestions. @Patrick : The article was good. I'm actually interested to implement the connectivity b/w the mobile & server usin... [more]

Thank you for all the suggestions. @Patrick : The article was good. I'm actually interested to implement the connectivity b/w the mobile & server using an app running on both systems(that reduces my effort in creating USB/Bluetooth data transfer ). @Aditya : I'm not sure about PcRemote as it is based on wifi hotspot(I use XP, which has lots of issues with this). I found WO Webcam working well with USB. @Sandeep : I'll definitely try this as it is USB, might work for me. Does it support audio for you? Or is it only video?

Following
Question:
Open Registration of two (possibly high-dimensional) probability distributions.
I am dealing with finding the affine transformation between two probability distribution functions in a dimensional space greater than 3 (e.g. 6). I d... [more]

I am dealing with finding the affine transformation between two probability distribution functions in a dimensional space greater than 3 (e.g. 6). I don't have the pdf shape. I have just samples from the two (unknown) distributions. The wo distributions have a different number of samples. I came up with a method that looks reasonable, works pretty well, and it's not that computationally expensive, but I don't know the state of the art. It is highly probable that someone already proposed a very similar method. But I don't know where to search. Can someone help me? Some bibliographical reference?

By Carlo Gatta · Autonomous University of Barcelona

Following
Answer added to:
3 What do we exactly mean by feature dissimilarity measure?
By Subrajeet Mohapatra · National Institute of Technology Rourkela

Mohammed Korayem · Indiana University Bloomington

You might check these papers: 1) S.-H. Cha. Comprehensive survey on distance/similarity measures between probability density functions. Inter- nationa... [more]

You might check these papers: 1) S.-H. Cha. Comprehensive survey on distance/similarity measures between probability density functions. Inter- national Journal of Mathematical Models and Methods in Applied Sciences, 2) A. Strehl, J. Ghosh, and R. Mooney. Impact of similarity measures on web-page clustering. In Workshop on Artificial Intelligence for Web Search (AAAI 2000), pages 58–64, 2000.

Following
Answer added to:
11 I am confused as to whether the eyes and particularly the iris remains the same at growing age? As I am working on Age invariant face recognition.
By Dulari Bhatt · Nirma University

Gary Feierbach · Independent Researcher

If you are close enough to scan the iris then you can probably can scan the retina which is unique. People can wear colored contact lenses making iris... [more]

If you are close enough to scan the iris then you can probably can scan the retina which is unique. People can wear colored contact lenses making iris patterns difficult to see and also distorting it. There is also the distance between the eyes and the 3D line patterns connecting the anterior notch of the ears, to the septum of the nose. These are dimensions and angles that shouldn't change.

Following

About Computer Vision

For discussion on techniques for aqcuiring and analysing images and other high dimensional data in order to produce information.

Topic Followers (14192) See all

Anshu Khandelwal

National Brain Research Centre
Shaban Bakari

Nizhny Novgorod State University
João Ferreira Nunes

Instituto Politécnico de Viana do Castelo
Pranjal Sahu

IIT Kharagpur
David Markus

Saddleback Geosolutions
Uzma Khattak

Abdul Wali Khan University Mardan
Aditya Rizki Yudiantika

Gadjah Mada University
Malik Mallem

Université d'Évry-Val-d'Essonne
Pedram Azad

Karlsruhe Institute of Technology
Ong Chin Ann

Swinburne University of Technology

Computer Vision

About Computer Vision

Topic Followers (14192) See all

Log in