AbstractWe investigate the problem of pedestrian detection in still images. Sliding window classifiers, notably using the Histogram-of-Gradient (HOG) features proposed by Dalal and Triggs are the state-of-the-art for this task, and we base our method on this approach. We propose a novel eature

extracti on scheme which computes implicit ‘soft egmentations’ of image regions into oreground/background. The method yields tronger object/background edges than grayscale

gradient alone, suppresses textural and shading variations,and captures local coherence of object appearance.-AbstractWe investigate the problem of pedestrian detection in still images. Sliding window classifiers, notably using the Histogram-of-Gradient (HOG) features proposed by Dalal and Triggs are the state-of-the-art for this task, and we base our method on this approach. We propose a novel　eature

extracti on scheme which computes implicit ‘soft egmentations’ of image regions into　oreground/background. The method yields　tronger object/background edges than grayscale

gradient alone, suppresses textural and shading variations,and captures local coherence of object appearance.