CSC669 - Selected Topics in Image Processing and Pattern Recognition

This course is focussed on enabling the students to do research on advanced topics of Image Processing, Pattern Recognition and related areas. After taking the course, the students must  

• understand the advanced level concepts of Deep Learning
• be able to apply Deep Learning to Image Processing and Pattern Recognition tasks such image denoising, image segmentation, classification and recognition.
• be able to apply the learned techniques to solve real life problems in areas like medical diagnosis, security, biometrics etc.

Topics (tentative): not necessarily in the order shown

Lectures	Topic
2	Basics: Introduction to Digital Image Processing and Pattern Recognition, Intensity Transformations and Filtering
3	Background on Deep Learning: Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Stacked Autoencoders (SAE), Recurrent Neural Network (RNN)
3	Dominant deep CNN model architectures: AlexNet, VGGNet, ResNet, GoogleNet, DenseNet etc.
2	Hand Engineered Feature Extraction: Local Binary Pattern (LBP) and its variants Histogram of Oriented Gradients (HOG) Scale invariant feature transform (SIFT)
1	Deep CNN as feature extractor,  Transfer learning, Pretraining and fine tuning
1	Interpretation of deep CNN activations
1	Image denoising using deep CNN
1	Image Segmentation using deep CNN