[ros-users] Reinventing the wheel: training and evaluation of algorithms for objects detection in images

[Victor Eruhimov]

Hello Rodrigo, copy-pasting my response from OpenCV user group. 

I think it is a great initiative and I am looking forward to see such a framework. OpenCV has a lot of infrastructure on object detection that you might want to look at, such as cascade classifier training, several machine learning classifiers and feature extraction methods. I really want to see OpenCV as a library that makes it easy to create an object detection application. OpenCV 2.2 will have some of what you call "install, run, see the graphs coming out" functionality such as a sample running bag-of-words object class detection on PASCAL VOC datasets. This framework can be used for generic object detector training too (PASCAL data structures + OpenCV methods for reading them). 

Having said that, I have to admit that OpenCV does not have a generic framework for training a detector for any type of object. And here is why: the problem is not solved yet. There are a few problems that are solved up to quality accepted by industry, face detection is one of the famous. If you look at the ROC curves from PASCAL VOC papers, a lot of them are far on the research side. OpenCV 2.2 will have two methods for detecting object classes: latent SVM and color descriptors with bag of words. Even if we try now to join them into a single framework, this will be quite hard -- the algorithms are too different. But since this is still a research problem, both approaches might change a lot over a couple of years and this will render our efforts useless. 

So my suggestion is to select the bits and pieces of object detection that are likely to survive and implement a framework for them. I already mentioned VOC datasets -- it is very important to have a data structure definition to work with. Local descriptors will be put in such a framework in OpenCV 2.2, this is a good example of "stable" functionality in my opinion.

Best Regards, Victor

On Nov 22, 2010, at 12:09 AM, Rodrigo Benenson wrote:

> Hello all.
> 
> 
> 
> I'm contacting you because I am considering starting a new open source project to solve a specific problem: training and evaluating objects detection algorithms. 
> 
> Hundreds of students have been there before: "I want to create a program that detects objects in images".
> 
> 
> 
> They choose a dataset for training (e.g. INRIA pedestrians), a feature descriptor (e.g. HOG), a machine learning method (e.g. linear SVM), and then, they write the code to get it all together.
> 
> In the best case they will take bits and pieces from multiple places and spend a few weeks on the glue code. In the worst case they will spend months reimplementing existing methods.
> 
> 
> 
> 
> It is time to stop the madness.
> Training detectors for objects detection in images is a specific and well defined problem. 
> It is time to share our effort and build a reference open source tool to solve this common problem.
> 
> 
> 
> We should have an open source tool that provides all the common bits, the glue and allows us to focus on what we really care: the algorithms.
> 
> In some sense OpenCv 2.2 helps a lot to the task, however OpenCv is aimed to be a generic library not a specific application framework. In that sense it will never provide the desired "install, run, see the graphs coming out" experience.
> 
> 
> 
> Also ROS.org helps a lot the task, by providing a generic framework to create and exchange software modules, along with standard tools for messages passing, data storage and exploration. However this framework by itself has a non negligible learning curve and it is unfamiliar to anyone outside the robotics community.
> 
> 
> 
> 
> I currently have my own idea of how things could be. However before creating "yet one more framework" I would like to have your input on the topic.
> 
> 
> 
> I have created a short form to collect your opinions. I would be very glad if you could help me go in the right direction by giving your input.
> 
> https://spreadsheets.google.com/viewform?formkey=dFFzaDlLM1liVGNOS2FENnhrc1VWckE6MQ
> 
> 
> 
> The form is anonymous and the results are public. 
> 
> Based your opinions and ideas I will do my best to move forward an usable open source solution.
> Further information will be posted at
> https://wave.google.com/wave/waveref/googlewave.com/w+yH-HOCb6H 
> 
> 
> 
> 
> Best regards,
> rodrigo benenson phd.
> 
> ps: If you are interested do not hesitate to send me a message. You can contact me via github as "rodrigob". 
> 
> 
> 
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20101123/b31448dc/attachment-0003.html>