摘要
人脸检测系统应用在嵌入式环境中需满足多种约束,高计算密集性、控制密集性是实时实现困难的主要原因.文中提出一种基于名为"REMUS-Ⅱ"的粗粒度动态可重构架构的人脸检测系统,把层叠型AdaBoost检测算法划分成多个非连续子任务,通过邮箱通信调度、配置流和数据流优化方法来提高指令级并行度和任务级并行度.实验结果表明,检测分辨率为640×480的图片可获得17帧/s的平均检测速度,正面人脸检测率保持在95%以上.在TSMC 65 nm CMOS工艺、200 MHz工作频率下,REMUS-Ⅱ面积约为24 mm^2,功率约为194 mW.
Face detection system needs to meet a variety of constraints in embedded environments, but the high computational/control intensive features make the real-time implementation difficult. This paper presents a face detection system based on a dynamical coarse-grained partially reconfigurable platform called "REMUS- Ir'. The cascade AdaBoost-based detection algorithm is divided into several non-consecutive sub-tasks. Mailbox scheduling, configuration flow and data flow optimization methods improve the instruction-level and task-level parallelism. Experiment results show that this approach with a 200 MHz clock can process about 17 frames per second on 640×480 images. Its detection rate is over area is about 24 mm2 in TSMC's 65 nm logic process. 95%. The system consumes about 194 mW, and its
出处
《应用科学学报》
EI
CAS
CSCD
北大核心
2012年第3期299-305,共7页
Journal of Applied Sciences
基金
国家“863”高技术研究发展计划基金(No.2009AA011700)
江苏省高校“青蓝工程”项目基金资助