Skip to content

README_SignLanguageRecognition.md

Latest commit

 

History

History
77 lines (51 loc) · 2.06 KB

README_SignLanguageRecognition.md

File metadata and controls

77 lines (51 loc) · 2.06 KB

🤟 Sign Language Recognition using Mediapipe & KNN

A real-time vision-based Sign Language Recognition system that converts American Sign Language (ASL) hand gestures into text and speech. Built using Python, MediaPipe, OpenCV, and K-Nearest Neighbors (KNN) algorithm.

🧠 Project Abstract

This project aims to bridge the communication gap between speech/hearing-impaired individuals and the general public. It recognizes static hand gestures representing A–Z alphabets in ASL and converts them into text on screen and speech using a Text-to-Speech engine.

⚙️ Features

  • Real-time hand gesture recognition using webcam
  • ASL A-Z alphabet detection
  • Word formation using continuous gestures
  • Text-to-Speech output of recognized words
  • Erase functionality to correct typed words

🛠️ Tech Stack

  • Python
  • MediaPipe
  • OpenCV
  • NumPy
  • Pandas
  • KNN Algorithm
  • pyttsx3 (TTS)
  • CSV for dataset handling

🚀 How It Works

  1. Image Acquisition: Captures hand gestures using webcam.
  2. Landmark Detection: Uses MediaPipe to extract 21 hand landmarks (x, y).
  3. Feature Engineering: Calculates 20 Euclidean distances from wrist to each point.
  4. Classification: Applies KNN algorithm with K=8 to classify the gesture.
  5. Output Display:
    • Displays detected alphabet on-screen (Char)
    • Forms a word (String)
    • Speaks out word using TTS
  6. Erase Gesture: Erases last character if Erase gesture is detected.

📈 Dataset Preparation

  • 40 images per alphabet (A–Z) captured using webcam.
  • Each image processed to extract 21 landmark points using MediaPipe.
  • Saved to .csv using landmarkextract.py.
  • Combined into datasets.csv and labeled accordingly.

📊 Algorithm Used

K-Nearest Neighbors (KNN)

  • K = 8
  • Classifies gesture based on nearest 8 vectors
  • Chosen for its simplicity and real-time performance

🔊 Additional Functionalities

  • Text-to-Speech using pyttsx3
  • Erase Gesture to remove last character
  • String Formation for words like “HOW ARE YOU”