This course covers the fundamentals of image processing and computer vision, along with core deep learning models such as feedforward networks, CNNs, RNNs, and transformers. It also introduces generative AI and recent trends, including explainability, zero-shot, few-shot, and self-supervised learning, while providing hands-on experience in implementing basic, deep learning-based, and advanced computer vision tasks. This course is also offered as an elective for PG-diploma students.

Lectures-Tutorials-Practicals-Credits: 4-0-0-4

Week 1: Image representation; Filtering; Frequency domain; Sampling.

Week 2: Edge detection; Feature extraction; Scale-space; SIFT; Segmentation; Visual perception.

Week 3: Neural networks; Backpropagation; Optimization; Regularization; Training improvements.

Week 4: Introduction to convolutional neural networks (CNNs); Backpropagation in CNNs; CNN architectures for image classification (AlexNet, VGG).

Week 5: CNN evolution; Modern architectures; Fine-tuning; Visualization.

Week 6: Object detection; Two-stage and single-stage models; Image segmentation.

Week 7: Introduction to recurrent neural networks (RNNs); Backpropagation through time; LSTM and GRU architectures; Video understanding using CNNs and RNNs.

Week 8: Attention in vision; Image captioning; VQA and dialog; Transformers.

Week 9: Vision transformers; Transformer-based object detection; Transformer-based image segmentation.

Week 10: Deep Generative Models: Review of (Popular) Deep Generative Models: GANs, VAEs; Other Generative Models: PixelRNNs, NADE, Normalizing Flows, etc

Week 11: Diffusion models and DDPMs; Denoising diffusion processes; Classifier-based and classifier-free guidance techniques.

Week 12: Contrastive and self-supervised learning; Vision-language models; Multimodal LLMs; Conclusion.