PoseCall
Gesture-controlled video calling application
Overview
PoseCall is a next-generation video communication platform that enables gesture-based interactions alongside traditional calling. Users can control calls through hand gestures, creating a more intuitive and engaging communication experience.
Problem
Standard video calling feels outdated. Users want richer, more expressive ways to communicate beyond just audio/video. There's potential for gesture recognition to enhance how we interact during calls.
Solution
PoseCall combines:
- Real-time video streaming with WebRTC
- Pose/gesture recognition using computer vision
- Gesture-based call controls and interactions
- Low-latency, reliable communication backend
Technical Architecture
Frontend:
- WebRTC for peer-to-peer video streaming
- Real-time pose detection and gesture recognition
- Interactive UI for gesture visualization
Backend Services:
- Signaling server for WebRTC connection management
- Gesture processing and normalization
- Session management and reliability
Real-time Processing:
- Low-latency gesture detection
- Concurrent connection handling
- Reliable media streaming
Key Achievements
- Demonstrates full-stack thinking (frontend + backend)
- Handles real-time media and pose data processing
- Low-latency architecture for responsive user experience
- Integration of computer vision with communication systems
Tech Stack
JavaScript
WebRTC
Computer Vision
Real-time Processing
Node.js
React