电子测量与仪器学报

孙锐,王慧慧,叶子豪.融合深度感知特征与核极限学习机的行人检测[J].电子测量与仪器学报,2019,33(2):39-47

融合深度感知特征与核极限学习机的行人检测

Pedestrian detection based on combining depth perception features with kernel extreme learning machine

DOI：

英文关键词:pedestrian detection DAGnet network kernel extreme learning machine saliency detection

基金项目:国家自然科学基金(61471154)、安徽省科技攻关科技强警项目(170d0802181)资助

作者	单位
孙锐	1.合肥工业大学计算机与信息学院,2.工业安全与应急技术安徽省重点实验室
王慧慧	1.合肥工业大学计算机与信息学院,2.工业安全与应急技术安徽省重点实验室
叶子豪	1.合肥工业大学计算机与信息学院,2.工业安全与应急技术安徽省重点实验室

Author	Institution
Sun Rui	1.School of Computer and Information, Hefei University of Technology,2. Anhui Province Key Laboratory of Industry Safety and Emergency Technology
Wang Huihui	1.School of Computer and Information, Hefei University of Technology,2. Anhui Province Key Laboratory of Industry Safety and Emergency Technology
Ye Zihao	1.School of Computer and Information, Hefei University of Technology,2. Anhui Province Key Laboratory of Industry Safety and Emergency Technology

摘要点击次数: 63

全文下载次数: 1

中文摘要:

行人检测在辅助驾驶和交通监测等方面有着广泛的应用，一直是计算机视觉领域中的研究热点和难点。传统特征提取方法对处在复杂环境中的行人难以有效地捕获具有区分度的特征信息。而目前流行的卷积神经网络因BP算法易陷入局部极小值，对泛化性能有所影响，且随着网络层的增加，一些显著特征信息逐层递减。针对上述问题，提出了融合深度感知特征与核极限学习机的行人检测算法。首先在CNN结构的基础上分两个阶段将前层特征与深层次特征融合后，送入后续层学习，构造一个DAGnet网络。随后采用实时性高，泛化能力强的核极限学习机对所得深度特征信息进行分类，并采用K 折交叉验证进行参数寻优；检测阶段，在DAGnet网络学习到的特征图上采用基于图论的显著性分析算法（GBVS），快速标注测试图像中行人的区域，然后在显著区域利用滑动窗口检测行人的精确位置。实验证明，所提算法在INRIA数据集和Caltech数据集的正检率均高于90%，在保证精度的情况下检测速度也得到明显提高。

英文摘要:

Due to the popularity and difficulty of research in the field of computer vision, pedestrian detection has been widely used in auxiliary driving and traffic monitoring. Traditional feature extraction method for pedestrians in complex environment is difficult to effectively capture the distinct characteristics information. And convolutional neural network which is popular at present has some influence on generalization performance because BP algorithm is easy to fall into local minimum value,and with the increase of the network layer, some significant feature information is decreasing layer by layer. In view of the above problems, this paper proposes a pedestrian detection algorithm combining deep sensing features with kernel extreme learning machines. Firstly, on the basis of CNN structure, the front layer features and the deep layer features are fused in two stages, and then sent to the follow up layer for learning, the directed acyclic graph network(DAGnet) network is constructed. Then, the depth feature information is classified by the kernel extreme learning machine with high real time performance and strong generalization ability, and the parameters are optimized by K fold cross validation;In the detection phase, graph based visual saliency(GBVS) saliency detection algorithm is used on the feature map learned by the DAGnet network to quickly mark the pedestrian area in the test image, and then sliding window is used to identify the precise position of the pedestrian in the salient area.The experimental results show that the positive detection rate on the INRIA data set and the Caltech data set is higher than 90%,and the detection speed is improved significantly if the accuracy is guaranteed.

查看全文查看/发表评论下载PDF阅读器