Mobile phone use behavior detection method based on target detection

Abstract

The invention provides a mobile phone use behavior detection method based on target detection, and relates to the technical field of target detection. The method comprises the following steps: firstly, inputting a video or an image of a to-be-detected mobile phone use behavior into a human hand target detection model, then detecting whether a human hand exists in the video or the image, storing a picture with the human hand into a human hand picture data set, and outputting the picture to a mobile phone classifier model; using the trained HOG + SVM model as a mobile phone classifier model to carry out binary classification on the human hand picture, judging whether the human hand picture is a human hand picture with a mobile phone or not, and outputting the human hand picture in three output modes; manually marking pictures in the human hand picture data set, deleting an irrelevant data set, renaming the data set and generating a configuration file; and finally, training and testing an HOG + SVM classifier model through the processed human hand data set and the configuration file to realize detection of mobile phone using behaviors. The method combines a candidate region-based target detection technology and an end-to-end-based target detection technology, and gives consideration to the detection accuracy and efficiency.

Description

technical field [0001] The invention relates to the technical field of target detection, in particular to a method for detecting the behavior of using a mobile phone based on target detection. Background technique [0002] At present, the detection of human hands in videos mainly uses the object detection technology in computer vision. Target detection technology is mainly divided into traditional target detection and current deep learning target detection. The traditional methods for human target detection mainly use context detection, HOG+SVM human hand detection and detection methods based on color features. The deep learning framework is mainly divided into two categories: two-stage target detection algorithms and one-stage target detection algorithms. The two-stage is to first generate a series of candidate boxes as samples by the algorithm, and then classify the samples through the convolutional neural network, such as R-CNN, Fast R-CNN, Faster R-CNN, etc. In the fi...

AI Technical Summary

Problems solved by technology

The most widely used mobile phone detection is the camera to detect whether the driver has made a phone call while driving, and the relatively mature algorithm is also based on face recognition technology for manual mobile phone detection. If the face is blocked, it cannot be detected, so Cannot be used in other scenarios

In addition, there are fewer datasets of people carrying mobile phones or objects, and the public datasets are only datasets of people.

Images