中文摘要：

针对经典行人检测算法（HOG＋SVM algorithm）因滑动窗口滑动次数过多引起的计算量过大问题，提出一种基于显著区域的行人检测算法。把提取的显著度和原图结合得出有效图，实现由整张图像行人检测到局部有效图行人检测的转变；用贝叶斯准则将有效区域和基于协方差的行人检测有机结合，达到在降低计算量的同时提高检测准确率的效果。在公开数据集INRIA上的实验结果表明，该算法降低了计算量，明显改善了误检率。

英文摘要：

Classic algorithms of pedestrian detection usually locate the latent position via sliding window techniques which resize the matching window and/or original images at different scales and scan the image.However,there are two main drawbacks in this method.First,resizing at a fix rate can not search through the whole scale space,resulting in the failure of inaccurate obj ect location.Second,resizing and scanning at various scales is usually time-consuming,which is improper for practical applications. To conquer these difficulties,a pedestrian detection method based on the salient information was proposed.The salient detection model and the traditional covariance matrix descriptor were combined in a Bayesian framework to detect pedestrians in the still image.At last,the efficiency of the proposed approach was compared with state-of-the-art results,which was demonstrated on the public INRIA dataset.