SiliconIntelligence

1. Introduction

The term Perception Processing encompasses processor support for technologies that can enable computers to perceive the world the way we humans do with our sensory faculties. It targets areas like object detection, recognition and tracking, speech and gesture recognition and multimodal abilities like lip reading to support speech recognition. The applications for perception processing are both immense and diverse. More and more computing devices are being invisibly embedded into our living environment, and we notice their existence only when they cease to serve us. For this fledgling computing fabric to develop into tomorrow's ubiquitous computing environment, the primary means of interacting with it should be human friendly ones like speech and gesture. Future mobile embedded environments need to support sophisticated applications such as speech recognition, visual feature recognition, secure wireless networking, and general media processing. Work environments from the board room to the garage will eventually feature human friendly and hands free interfaces to the computers embedded into those environments. Perception prosthetics are an important application too. Devices that listen to speech and then project text on a heads up display worn by a deaf person or an intelligent camera that gives audio cues like ``Vehicle approaching,'' ``Stairs 10 feet ahead'' to a blind wearer are of particular interest. Another important application area is robotics - the opportunity to outfit both manned and autonomous vehicles, industrial and house hold robots and even machine tools with tireless vision presents boundless opportunity. Other areas that could benefit from perception processing include automated surveillance, translation of speech and a variety of assistive technologies.



Subsections

Binu Mathew