Multi-target detection method based on a cascade hourglass neural network

Abstract

The invention provides a multi-target detection method based on a cascaded hourglass neural network, and aims to solve the technical problems that an existing detection method is too slow in speed anda small target is difficult to identify. The method comprises the following steps: step 1, collecting a training sample; step 2, building a deep learning framework and a backbone network cascade hourglass network of target detection; step 3, designing a label of the training sample as a confidence coefficient heat map; step 4, designing a loss function of the cascade hourglass network to optimizethe loss function; step 5, training the cascaded hourglass network to obtain a detection model; And step 6, multi-target detection. The method has the beneficial effects that various types of targetscan be quickly and accurately identified, and the small target identification capability is improved.

Description

technical field [0001] The invention relates to the technical field of multi-target detection, in particular to a multi-target detection method based on a cascaded hourglass neural network. Background technique [0002] Multi-target detection is an important direction in the field of computer vision. The main task is to locate the target of interest from the image and judge the specific category of each target. It has been widely used in medical target detection, intelligent video surveillance, vehicle automatic driving, pedestrian detection, traffic flow statistics, etc. Traditional methods use directional gradient histograms, local binary pattern features, etc. to extract image features, and then use support vector machines, random forests, and neural networks to classify. But it has two disadvantages of slow speed and low precision. [0003] Convolutional neural networks have achieved large-scale success in images in recent years. Girshick proposed that RCNN and Fast-R...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a multi-target detection method based on cascaded hourglass neural network to solve the problems of poor effect and slow speed when the existing detection model is applied to identify small targets

Images