Convolutional neural network-based target detection method and system

Abstract

The invention discloses a convolutional neural network-based target detection method and system. The method comprises the steps of constructing a data set of a detected target; dividing the image datain the data set of the detected target into a training set, a test set and a verification set according to a preset proportion, and marking images in the training set, the test set and the verification set; constructing a network structure of a convolutional neural network model, wherein the convolutional neural network model adopts different feature scales to predict an object; loading a training set into the convolutional neural network model for training; in the training process, a verification set is loaded, and parameters of the convolutional neural network model are optimized through amulti-verification method; carrying out performance test on the convolutional neural network model through the test set, and detecting the generalization capability of the convolutional neural networkmodel; and carrying out target recognition is carried out by adopting a convolutional neural network model with the generalization capability meeting requirements. The convolutional neural network model obtained through training can quickly and accurately identify small targets, compact and dense targets or highly overlapped targets in a shopping mall.

Description

technical field [0001] The present invention relates to the field of target recognition, in particular to a method and system for target detection based on a convolutional neural network. Background technique [0002] In recent years, with the development of deep learning of artificial intelligence technology, especially after the convolutional neural network technology was proposed, the method of target detection has been continuously updated, and the detection effect has also achieved a qualitative leap. [0003] In view of the rapid development of the existing target detection technology, it has also been widely used in large shopping malls. For example, in shopping malls, it is necessary to carry out unmanned retail identification of commodity categories, face recognition matching at the entrance, and so on. Therefore, in large shopping malls, target detection technology has very good development prospects. In the existing technology, given an input image, an object or...

AI Technical Summary

Problems solved by technology

[0004] In the existing target detection technology, the detection speed can roughly meet the demand, but there are still big defects in the accuracy and recognition effect

In the existing target detection technology, the detection effect of a single object is better, but if it is necessary to detect multiple targets that are densely packed or highly overlapping targets, or when the detected object accounts for a small proportion of the overall figure, it cannot be accurately and quickly achieved. Expected target detection effect

This defect is especially serious in the scene application of large shopping malls. The products are of different sizes and placed densely, and there are a large number of mobile people. The existing image detection technology cannot meet the application in this scene.

Moreover, the generalization ability of the existing target detection technology is weak, and the object classification cannot achieve the desired effect.

If the same type of object has a new and uncommon aspect ratio and other situations, it will not be able to complete the ideal detection effect

Images