Target detection method based on improved Mixed Pooling-YOLOV3

Abstract

The invention discloses a target detection method based on improved Mixed Pooling-YOLOV3, and the method carries out the design of a target detection algorithm based on the idea of regression, and achieves the multi-scale and multi-label classification. Based on the defects of a current one-stage target detection method YOLOV3, a DPM network based on a Darknet-53 structure is designed as a featureextractor; and secondly, an original loss function is reconstructed, and meanwhile, parameters of the original loss function are optimized, so the detection precision is effectively improved, and convergence is quicker. The DMP framework gives consideration to the complexity of the network and the accuracy of detection; compared with a common target detection feature extraction network VGG-16, the method has the advantages that the model calculation amount is reduced, the latest progress of computer vision is introduced into the field of target detection, the detection precision and the detection speed are good, and meanwhile, the method has better performance and popularization and application prospects.

Description

technical field [0001] The present invention relates to the technical fields of computer vision, deep learning technology and target detection, in particular to a target detection method based on improved Mixed Pooling-YOLOV3. Background technique [0002] As one of the most fundamental and challenging problems in computer vision, object detection has received tremendous attention in recent years. Object detection is a basic computer vision task, which provides basic information for semantic segmentation of image and video understanding, and can also be used to detect instances of specific categories of visual objects in digital images, so it has received extensive attention. The goal of object detection is to develop computational models and techniques that provide the basic information needed for computer vision applications: what is the object, and where is it? From the perspective of application, target detection can be divided into two research topics: "general object ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to integrate the ideas of "hybrid pooling" and "reconstruction loss function" into the network structure of YOLOV3 in view of the problems of low precision, unbalanced positive and negative samples, and overfitting of the one-stage target detection algorithm YOLOV3 In, a new DMP (Darknet based on Mixed Pooling) target detection network framework is proposed

Images