Object detection network design method based on image segmentation feature fusion

Abstract

The invention discloses a design method of a target detection network based on fused image segmentation features, which is effective for large-scale targets. Based on the general target detection framework Mask RCNN and image segmentation feature fusion, the target segmentation feature and a ResNet-101 convolutional network are integrated into the rpn module, an RoI Pooling module and an RoI Alignmodule, the experiments show that the method is effective for large targets, and the image segmentation algorithm for small targets can be improved completely if the image segmentation effect is ideal.

Description

technical field [0001] The invention belongs to the field of pedestrian detection methods, in particular to a target detection network design method for fusing image segmentation features. Background technique [0002] With the development of science and technology and the progress of the times, we have to admit that our way of life is also constantly changing. People's travel methods are constantly updated. Cars are the most widely used means of transportation in the contemporary environment. According to statistics from the Traffic Management Bureau of the Ministry of Public Security, as of the end of June 2017, the number of motor vehicles in the country reached 304 million, including 205 million cars. Colleagues The problem of traffic safety is particularly prominent. According to incomplete statistics, the number of people killed in traffic accidents in low- and middle-income countries has reached more than 90% of the total number of deaths in the world every year, but ...

AI Technical Summary

Problems solved by technology

Although the current research has been able to judge the images of pedestrians to a certain extent, there are still many problems that cannot be fully and accurately identified.

[0005] In the field of pedestrian detection, the early feature extraction mainly used HOG features, but because the HOG features are artificially designed, the feature extraction algorithm process is fixed, and only pedestrians can be better recognized when they maintain a standing posture. The idea of ​​fusion is to fuse HOG features with other image features, such as image segmentation features, image depth features and image edge features, etc.

Recently, convolutional neural networks have developed rapidly in the field of computer vision and have gradually replaced artificially designed features, but their performance still needs to be improved, and the idea of ​​feature fusion is still applicable here

Images