Abstract

Introduction: Communication with the hearing impaired ( deaf/mute) people is a great challenge in our society today; this can be attributed to the fact that their means of communication (Sign Language or hand gestures at a local level) requires an interpreter at every instance. Conversion of images to text as well as speech can be of great benefit to the non-hearing impaired and hearing impaired people (the deaf/mute) from circadian interaction with images. To effectively achieve this, a sign language (ASL – American Sign Language) image to text as well as speech conversion was aimed at in this research. Methodology: The techniques of image segmentation and feature detection played a crucial role in implementing this system. We formulate the interaction between image segmentation and object recognition in the framework of FAST and SURF algorithms. The system goes through various phases such as data capturing using KINECT sensor, image segmentation, feature detection and extraction from ROI, supervised and unsupervised classification of images with K-Nearest Neighbour (KNN)-algorithms and text-to-speech (TTS) conversion. The combination FAST and SURF with a KNN of 10 also showed that unsupervised learning classification could determine the best matched feature from the existing database. In turn, the best match was converted to text as well as speech. Result: The introduced system achieved a 78% accuracy of unsupervised feature learning. Conclusion: The success of this work can be attributed to the effective classification that has improved the unsupervised feature learning of different images. The pre-determination of the ROI of each image using SURF and FAST, has demonstrated the ability of the proposed algorithm to limit image modelling to relevant region within the image.

Keywords: Image and Speech processing, Text-to-Speech (TTS), Unsupervised Learning, and FAST and SURF algorithms