Human body image detection method based on HOG features

Abstract

The invention relates to the human body sensing technology, and provides a human body image detection method based on HOG features. The method focuses on the improvement of HOG feature extraction. In the detection process, firstly the HOG features of a target object in an image or video are extracted, and the extraction method comprises the steps that graying processing is carried out on an original image, a colorful image is converted into a gray level image, first-order gradient calculation is carried out on the image, an image region is divided into a plurality of cells, the gradient of each cell is calculated, statistics is carried out on a first-order gradient histogram of all pixels in each cell, and all the cells are normalized on blocks. Finally, the HOG features of all the blocks of a space are collected and detected to obtain the HOG features of the target object. After the HOG feature extraction of the target object is completed, a classifier with a judgment function is trained and used for judging the HOG features of the target object. The method is suitable for the field of human body sensing.

Description

technical field [0001] The invention relates to human body perception technology, in particular to a human body perception method based on HOG features. Background technique [0002] Object detection and recognition is an important research field in computer vision, which refers to the ability of computers to detect and recognize specific objects according to human thinking. Its application is extremely extensive, and fast and accurate object detection and recognition technology is an important part of modern information processing technology. Due to the rapid increase in the amount of information in recent years, we also urgently need suitable object detection and recognition technologies that allow people to find the information they need from a large amount of information. Image retrieval and text retrieval both fall into this category. In addition, object detection and recognition technology is also widely used in public security and traffic supervision systems. Face d...

AI Technical Summary

Problems solved by technology

[0004] HOG features are widely used. In current computer vision applications, HOG features are mainly used for pedestrian detection. Due to the fast calculation speed, there are defects in detection accuracy and other aspects.

Moreover, there are great changes in the body shape, posture, clothing and lighting of different pedestrians, and there are technical difficulties in complex backgrounds.