IOS-Based People Detection of Multi-object Detection System

In recent years, people detection and face recognition technology developed quickly, including people detection part there is considerable room for development. People detection can be used in vehicle auxiliary systems or monitoring systems to identify people by the camera to capture the image. However, the people detection accuracy and running speed is also a very important part, so this study presents an iOS-based people detection system, the proposed system of the people detection applied to mobile devices. This study will explore the accuracy and running speed for the HOG method of OpenCV on the iOS-based system and try to improve these problems. Through histogram of oriented gradient(HOG) method to turn images into grayscale, histogram equalization, dividing the computational domain and computing gradient feature vector and using the support vector machine(SVM) classifier to do classify, and then compare with INRIA database to determine whether a pedestrian and mark the location of pedestrians. Finally, the region of interest is added to reduce the amount of computation, so it can improve the performance of computing, analysis the accuracy and computational time of the experimentally result.