SVM Tutorials

Currently working on unknown action classification based on the output of SVM classifier for known action classes. Basic idea is how to segregate unknown actions from the known actions given the knowledge that we already have. Below is a link with tutorial with good SVM tutorials. http://svms.org/tutorials/...

Read more »

Experimental tasks

List of tasks (8/8/2012)    Verify the correctness of implementation on all processes of existing work Done usingKTH and Weizmanns action dataset Benchmark against VLFeat's BOW using Laptev's HOG-HOF descriptor List of different processes involved: STIP detection Number of interest points BOW formation Vector quantization Histogram binning Building of classifier ST Correlation...

Read more »

Computer vision resources

Links of various resources  1. Contains codes for Human Activity Recognition and Tracking http://www.cis.jhu.edu/~rizwanch/research.htm 2. Contains code for computing Histogram of oriented gradients used in Laptev CVPR 2008 paper http://lear.inrialpes.fr/people/triggs/src/ 3. Contains code for event detection http://www.fit.hcmus.edu.vn/~vdphong/Event%20Detection.htm 4. Links to various...

Read more »

List of important deadlines and administrative tasks

Deadlines  FG 2013 Shanghai (official: 15/ 9/ 2012) Finish coding (Results) - 14/8/2012 Paper - 21/8/2012  Content:  Detection of unknown action classes based on probabilities Ability to distinguish from trained action classes Includes results done using inter-cross-validation between different datasets e.g. Weizmann vs KTH VISAPP 2013 Spain (official: 11/9/2012) Finish coding...

Read more »