Projects

Cognitive Vision-Based Hand Gesture Recognition System for Media and Accessibility Control

Python, OpenCV, Google MediaPipe

Developed a two-handed gesture system that controls system volume, brightness, and media playback on a consumer CPU. This was achieved by incorporating geometry mappings and focusing only on the necessary physical actions of hand landmarks mapped to digital actions to reduce the user’s cognitive load. Average latency of 35 ms; ∼28–30 FPS in good lighting, ∼15–20 FPS in low light on consumer-grade CPUs.

ROS Robot – Controller Input

Raspberry Pi 3 Model B, Arduino, CamJam EduKit, ROS Noetic, Python

Configured teleoperation nodes and subscriber/publisher communication pipelines for a chassis robot built on Ubuntu OS and ROS Noetic. Integrated Arduino microcontrollers.

Pathfinding Visualization GUI

Python, Pygame

Implemented Dijkstra’s Algorithm, Breadth-First Search (BFS), and A* Search for graph traversal. All algorithms try to find the shortest path from one point to another.

Not3s – Cross-Platform Note-Taking Application

Flutter SDK, Dart

Developed a cross-platform note-taking application using the Flutter SDK. It uses Firebase Cloud Messaging and whisper_kit to enable data synchronization and accurate on-device dictation.