UnVoiced Sign AI

Project Architecture & Overview

Accessibility assistive system enabling deaf and mute individuals to communicate freely. Captures hand signs through computer vision, isolates skin tones using YCrCb color space thresholding, classifies gestures, and synthesizes natural audible speech in real time.

Key Engineering Milestones

• ✦Real-time Indian Sign Language alphabet and numeric gesture parsing
• ✦Robust YCrCb color masking adapting to varying ambient skin and light conditions
• ✦Integrated text-to-speech synthesis generating instant voice audio output
• ✦Low hardware requirement: Runs on budget dual-core laptops and Raspberry Pi