Animal Sign language recognition using MEMS

Mple dog’s express their feelings by oscillating his tail. Similarly cows express their feelings with his feet, neck and tail. Every feeling has its distinct body movements. So by placing this model on their communicating body part we transform the animal feelings into words and phrases. This model also helps the deaf and dumb community to communicate with others in oral langage. Prototype of sign language recognition consists of ADXL335 accelerometer interfaced with PIC micro controller 16F873A. The interfacing program is written in embedded ‘C’ language and it is compiled with Hi-tech compiler. The accelerometer data is processed in PC using neural network pattern recognition tool available in MATLAB. In this model we transformed six American Sign Language postures into words. We proposed two algorithms based on Euclidean distance metric and neural network pattern recognition tool with spline interpolation technique achieving an overall l efficiency of 80% and 83.3% respectively. Former algorithm is preferred because here we achieve an efficiency varying from 70% to 90% where as in Euclidean distance algorithm efficiency varies from 0% to 100% Keywords-ASL, Euclidean distance, nprtool, MEMS accelerometer, animal sign language, spline interpolation, interpolation technique