FPGA Implementation of Human Detection by HOG Features with AdaBoost
FPGA Implementation of Human Detection by HOG Features with AdaBoost
复制标题
DOI：
10.1587/transinf.e96.d.1676
复制
发表时间：
2013-08
期刊：
影响因子：
--
通讯作者：
Keisuke Dohi;Kazuhiro Negi;Yuichiro Shibata;K. Oguri
中科院分区：
文献类型：
--
作者：
Keisuke Dohi;Kazuhiro Negi;Yuichiro Shibata;K. Oguri
我们实现了无需外部存储的深度流水线的FPGA实现，包括HOG特征提取和AdaBoost分类。为了用紧凑的现场可编程门阵列构建我们的设计，我们引入了一些简化的算法和积极使用面向流的体系结构。我们给出了简化的定点方案和原始浮点方案在结果质量方面的比较结果，结果表明简化的方案对硬件实现的负面影响是有限的。我们的实验表明，在Xilinx Virtex-5 XC5VLX50上，我们的系统能够以高达112FPS的速度从640×480 VGA图像中检测出人体。关键词：方向梯度直方图，AdaBoost，人体检测，现场可编程门阵列
We implement external memory-free deep pipelined FPGA implementation including HOG feature extraction and AdaBoost classification. To construct our design by compact FPGA, we introduce some simplifications of the algorithm and aggressive use of stream oriented architectures. We present comparison results between our simplified fixed-point scheme and an original floating-point scheme in terms of quality of results, and the results suggest the negative impact of the simplified scheme for hardware implementation is limited. We empirically show that, our system is able to detect human from 640 × 480 VGA images at up to 112 FPS on a Xilinx Virtex-5 XC5VLX50 FPGA. key words: histogram of oriented gradients, AdaBoost, human detection, FPGA