Some visualization are not displayed here, so please check it open it on collab too.
RF, SVM
A vocabulary of 300 visual words is used to represent each image.
The SIFT approach for encoding image patches is not available on python openCV library anymore.
Therefore, the descriptors were extracted on MATLAB and analyzed on Python.
Kmean method was used for bagging the descriptors.
Coursework2
GAN, CGAN, DCGAN
About
Code for coursework1&2 of CS492A - Machine Learning for Computer Vision, KAIST