Motion segmentation using optical flow for pedestrian detection from moving vehicle

Abstract

This paper proposes a pedestrian detection method using optical flows analysis and Histogram of Oriented Gradients (HOG). Due to the time consuming problem in sliding window based, motion segmentation proposed based on optical flow analysis to localize the region of moving object. A moving object is extracted from the relative motion by segmenting the region representing the same optical flows after compensating the ego-motion of the camera. Two consecutive images are divided into grid cells 14x14 pixels, then tracking each cell in current frame to find corresponding cells in the next frame. At least using three corresponding cells, affine transformation is performed according to each corresponding cells in the consecutive images, so that conformed optical flows are extracted. The regions of moving object are detected as transformed objects are different from the previously registered background. Morphological process is applied to get the candidate human region. The HOG features are extracted on the candidate region and classified using linear Support Vector Machine (SVM). The HOG feature vectors are used as input of linear SVM to classify the given input into pedestrian/ non-pedestrian. The proposed method was tested in a moving vehicle and shown significant improvement compare with the original HOG.