Created by Suvam bag and Sourav Kulhare in the Machine Intelligence Lab in Rochester Institute of technology. -----Speacial thanks to P. Sconanner, S. Ali and M. Shah A 3D Dimentional SIFT Descriptor and its application, Computer Vision Lab, UCF.ACM MM 2007-----
#Introduction Video classification is one of the most difficult and important branches of computer vision. This project concentrated on classifying the KTH dataset using 3D SIFT. The dataset contains 6 different classes of human activity like walking, running, jogging, boxing, clapping, and waiving. The project can also be considered as an elementary atatempt on scene undertsanding whcih is one of the current research areas of computer vision and machine learning.
#Platform Matlab
#Description The experiments were conducted on a local machine and achieved 90% accuracy on certain classes after cross validation. The final paper has been uploaded for a detailed explanation.