Current Issue Cover
利用级联SVM的人体检测方法

李同治1, 丁晓青1, 王生进1(清华大学电子工程系,智能技术与系统国家重点实验室,北京 100084)

摘 要
从图像中检测出人体是计算机视觉应用中的关键步骤。通过一个由简到繁的级联线性SVM分类器将级联拒绝的机制与梯度方向直方图特征相结合,实现了一个准确和快速的人体检测器,整个检测器由级联的线性SVM分类器组成。实验结果表明,在保持Dalal算法检测准确性的同时,大幅的提高了检测速度,每秒平均可以处理12帧左右的320×240的图像。
关键词
Human Detection with a Coarse to fine Cascade Linear SVM

()

Abstract
Finding human in images is critical for several applications in computer visionWe combine the cascade of rejection approach with the Histograms of Oriented Gradient (HOG) to form a fast and precise human detectorThe detector consists of coarse to fine linear SVM classifiersOur experiments show that our method can process average 12 frames with 320×240 image per second,while maintaining the comparable accuracy to Dalals method
Keywords

订阅号|日报