A Reconfigurable Architecture for Rotation Invariant Multi-View Face Detection Based on a Novel Two-Stage Boosting Method

摘要

We present a reconfigurable architecture model for rotation invariant multi-view face detection based on a novel two-stageboosting method. A tree-structured detector hierarchy is designed to organize multiple detector nodes identifying pose rangesof faces. We propose a boosting algorithm for training the detector nodes. The strong classifier in each detector node is composedof multiple novelly designed two-stage weak classifiers. With a shared output space of multicomponents vector, each detectornode deals with the multidimensional binary classification problems. The design of the hardware architecture which fully exploitsthe spatial and temporal parallelism is introduced in detail. We also study the reconfiguration of the architecture for finding anappropriate tradeoff among the hardware implementation cost, the detection accuracy, and speed. Experiments on FPGA showthat high accuracy and marvelous speed are achieved compared with previous related works. The execution time speedups rangefrom 14.68 to 20.86 for images with size of 160 x 120 up to 800 x 600 when our FPGA design (98 MHz) is compared with softwaresolution on PC (Pentium 4 2.8 GHz).