Applications of Computer Vision and Deep Learning

RKMVERI CS/BDA | Spring 2026

Instructor: Jimut Bahan Pal
Credits: 3
Total Hours: 50 hrs

View Grade Sheet

Course Details

Attendance Policy: Credited students must maintain at least 75% attendance to receive a grade.

Comprehensive Curriculum

Commencing with foundational Python3, we advance toward pragmatic implementation using OpenCV and PyTorch, bridging the gap between theoretical math and deployment-ready applications.

Hands-on Learning

This tutorial-driven program blends theory with immersive tutorials. It is designed for those enthusiastic about applying classical and cutting-edge deep learning algorithms to real-world scenarios.

Research & Industry Focus

Participants attain mastery in designing models for classification, segmentation, and LLMs, critical skills for creating real-world applications in both industry and research domains.

Course Curriculum

  • • Overview & PyTorch Basics
  • • Theory: Softmax, Cross-Entropy
  • • Building Classification Pipelines
  • • DNN Parameter Calculation
  • • Variational Auto-Encoders (VAE)
  • • DL-based Segmentation Methods
  • • CNNs (DLNLP Guest Lecture)
  • • Transformers & BERT
  • • Transformer-based Models & Tricks
  • • Large Language Models (LLMs)
  • • LLM Training & Tuning
  • • LLM Reasoning & Evaluation
  • • Agentic LLMs
  • • Current Trends in AI

* We actively use Google Space for communication and slide sharing. Join the Google Space here.
* Find the grade sheet here: Grade Sheet

Schedule & Materials

Total Lecture Covered: 44.5 hrs
(Check attendance percentage: here)
Date Time Topic Materials
15-01-2026 7:00 - 8:30 PM (1.5 hours) Lec-0: Overview of the course
18-01-2026 8:30 - 12:00 AM (3.5 hours) Lec-1: Overview & General Introductions
25-01-2026 8:30 - 11:30 PM (3 hours) Lec-2: Basics of Pytorch and Activation functions in DNNs
01-02-2026 8:30 - 11:30 PM (3 hours) Lec-3: Perceptron, MLP and Pytorch Pipeline for classification
08-02-2026 Due 23rd-Feb-2026, 11:59 PM Assignment 1: XOR Classification (Theory + Practice)
08-02-2026 8:30 - 11:30 PM (3 hours) Lec-4: Theory of Likelihood, activations, gradient descent, and discussions on assignment
15-02-2026 8:30 - 11:30 PM (3 hours) Lec-5: Some theory of Backprop and Convolutional Neural Networks
22-02-2026 8:30 - 11:30 PM (3 hours) Lec-6: Parameter Calculation in Neural Networks
01-03-2026 8:30 - 11:30 PM (3 hours) Discussions and evaluations of Assignment-1
01-03-2026 Due 16th-March-2026, 11:59 PM Assignment 2: Lightweight CNNs and Explainable AI
08-03-2026 3:00 - 6:00 PM (3 hours) Lec-7: Variational Autoencoders
14-03-2026 1:15 - 4:15 PM (3 hours) Lec-8: Word2Vec Models, Sequence modelling and RNNs
15-03-2026 Talk by Vaibhab Rathore HiDISC: A Hyperbolic Framework for Domain Generalization with Generalized Category Discovery
15-03-2026 8:30 - 11:30 PM (3 hours) Lec-9: LSTMs and Attention-based models
21-03-2026 1:00 - 4:30 PM (3.5 hours) Lec-10: Transformers-1; Discussions and evaluations of Assignment-2
22-03-2026 Talk by Aniket Thomas GENIE: A Visual-Only Diffusion Framework for Task- Agnostic Image Transformation
22-03-2026 8:30 - 11:30 PM (3 hours) Discussions and evaluations of Assignment-2 & Talk
29-03-2026 8:30 - 11:30 PM (3 hours) Lec-11: Transformers-2; Discussions and evaluations of Assignment-2
29-03-2026 Due 13th-April-2026, 11:59 PM Assignment 3: Sentiment analysis using CNN
05-04-2026 Talk by Munish Monga (1.5 hrs) DuET: Dual Incremental Object Detection via Exemplar-Free Task Arithmetic (from Sony Research)
05-04-2026 Talk by Seshadri Mazumder (1.5 hrs) Towards Scalable Sign Production: Leveraging Co-Articulated Gloss Dictionary for Fluid Sign Synthesis (IIIT Hyderabad)