PROJECT TITLE:

An AdaBoost-Based Face Detection System Using Parallel Configurable Architecture With Optimized Computation - 2015

ABSTRACT:

With the event of image sensor technology, the AdaBoost-based mostly face detections are widely utilized in many monitoring sensor networks and mobile-camera-based applications. Fast face detection with high detection accuracy and low power consumption in such types of applications is very vital. Since the AdaBoost-primarily based face detection exhibits characteristics of information computation in dual direction and knowledge diversity, we propose an AdaBoost-based mostly face detection system using parallel configurable design with optimized computation. The architecture consists of parallel configurable arrays and 2-level shared memory systems. It achieves dual-direction-based mostly integral image computation that improves parallel processing potency and allows the subwindow adaptive cascade classification for information diversity, which any improves the detection potency in diverse face detection. Compared with the state-of-the-art works, this work achieves maximal performance of 30 frames/s at 1080p detection speed and extreme low power consumption.