OpenCV library is in case. Please, get more information through the link:

http://docs.opencv.org/modules/nonfree/doc/feature_detection.html

OpenCV library is in case. Please, get more information through the link:

http://docs.opencv.org/modules/nonfree/doc/feature_detection.html

For example: http://www.robots.ox.ac.uk/~vedaldi/code/siftpp.html

What about siftWin32.exe? Too slow?

Did you give a try to scikit-learn and scikit-image python modules ? These 2 modules have very nice feature for this kind of work. I went to EuroSciPy 2014 in Cambridge UK and there was a nice simple image recognition tutorial example with this two packages. Was pretty straight forward. The tutorial may still be at:

https://github.com/GaelVaroquaux/sklearn_europython_2014

Hope this helps.

Best regards,

Sylvain.

https://github.com/GaelVaroquaux/sklearn_europython_2014

You could also use VLFeat:

http://www.vlfeat.org/overview/sift.html

Regards,

Georg.

This guy has extremely good toolbox:

http://koen.me/research/colordescriptors/

It supports many types of SIFT (different color spaces), batch processing and GPU acceleration.

I add SIFTGPU ref. to the list

http://cs.unc.edu/~ccwu/siftgpu/

Try OpenCV

OpenCV is what you are looking for!

http://opencv.org/about.html

OpenCV!!

Hi! I'm student and i'am working with opencv for Android devices i found this implementations (SURF and SIFT) in this page: http://web.guohuiwang.com/technical-notes/sift_surf_opencv_android

i hope it can help you!

See ya!

http://koen.me/research/colordescriptors/

Thank you all so much, Amaury Lendasse, Vladimir R. Tuz, Rafal, Pierfrancesco, Volker etc.!

I am now trying OpenCV and Scikit-learn. They are suitable for my task. Also, I have found another machine learning tool box called "SHOGUN"(http://www.shogun-toolbox.org/), which beats many other toolboxes, e.g. scikit-learning, in many aspects (see their comparison report at: http://www.shogun-toolbox.org/page/features/).

Happy Sharing!

He Zhang

users.ics.aalto.fi/hezhang/ 

Hi!

Apart from SIFT or SURF, the OpenCV libraries include other interesting feature extraction algorithms that you can use for your proposal. In fact, you can use binary descriptors such as BRIEF, BRISK, ORB or FREAK. They are implemented in the OpenCV C++ API:

http://docs.opencv.org/modules/features2/doc/feature_detection_and_description.html

Binary descriptors are faster to compute with respect to other kind of vector-based descriptors such as SIFT and SURF. Moreover, they obtain remarkable results in image recognition problems such as the described in your question. I attach you two papers where we apply binary descriptors for place recognition. In this case, we obtain a high effectiveness with a low computational cost by using different binary descriptors.

Good luck with your research!

Conference Paper Fast and Effective Visual Place Recognition using Binary Cod...

Conference Paper Bidirectional Loop Closure Detection on Panoramas for Visual...

Thank you Roberto! I will cite your paper in my future research.

maybe you can also combine a feature extraction algorithm with a tracking one - http://personal.ee.surrey.ac.uk/Personal/Z.Kalal/tld.html

I think the following link is very helpful:

http://scikit-image.org/docs/dev/api/skimage.feature.html

It contains the most used image feature extraction methods (DAISY, HOG, LBP, MB-LBP, BRIEF, Haar-like, etc.) with examples.

https://github.com/topics/feature-extraction?l=c%2B%2B

https://github.com/topics/feature-extraction?l=python

https://pysource.com/2018/03/21/feature-detection-sift-surf-obr-opencv-3-4-with-python-3-tutorial-25/

https://github.com/chewbacca89/OpenCV-with-Python