Machines That Understand Video – 1st Computer Vision Virtual  Showcase  (CVVS -25) 

Course Code : 23AID301 Computer Vision (2- 0- 2-3)                                                                           S5 B.Tech. AS &AI                                     Faculty Information :  Dr. Don S 

This showcase presents the final course projects developed in the Computer Vision module, demonstrating the application of AI to understand, describe, and interpret visual information. Students explored core techniques such as image enhancement, feature extraction, segmentation, and audio-visual processing. Building on these fundamentals, the projects progressed into advanced tasks, including scene detection for understanding context, scene captioning and caption generation for producing natural language descriptions of visual content, and event proposal for identifying and localizing key actions within videos. Together, these works highlight how modern AI systems transform raw visual data into meaningful stories and insights. 

Using AI Tools for creating Video Presentations: Why and How 

Student groups use AI tools to create video presentations because these tools simplify content creation, automatically generate visuals aligned with computer vision concepts, and save time while ensuring accuracy and engagement. They typically use AI platforms by inputting key topics or scripts, which the tools then transform into structured visuals, animations, and narrated sequences for a cohesive presentation. 

Weekly  Activity

Expected Output

1 Project Orientation + Topic Finalization

Final project topic selected per group. Supervisor assigns reading material.

2 Literature Survey & Tool Familiarization

Submit 2-page literature survey. Explore tools/libraries (OpenCV, PyTorch, etc.)

3 Dataset Selection & Preprocessing

Download dataset, annotate/clean data, document the structure

4 Model/Method Exploration

Implement baseline model/approach from literature (even if small-scale)

5 Basic Functionality Implementation

Each group shows progress: feature extraction, scene labeler, etc.

6 Midway Review 1 (Mini Presentation) 3 marks

Present architecture, initial results, difficulties

7 Component Tuning & Evaluation

Try tuning hyper parameters, test on validation set

8 Advanced Model Integration / Optimization

Integrate improved methods (e.g., attention in captioning, ResNet for scenes)

9 Result Analysis & Visualization

Show result samples, evaluation metrics, errors

10 Midway Review 2 (Demo Focus)

Working demo of partial pipeline (e.g., event proposal from a 30s video)

11 Refinement & Documentation

Refactor code, create diagrams, update report

12 Final Demo Preparation  5 marks

Finalize slides, GitHub repo, test demo end-to-end

13 Project Presentation & Viva  2 marks

10-min group presentation + Viva + Report submission

Paper 8 : Enhancement of Lung CT Images for Improved Diagnostic Clarity