Neural network training and image recognition method and device, equipment and storage medium

Abstract

The invention relates to a neural network training and image recognition method and device, equipment and a storage medium. The training method comprises the following steps: extracting a neural network model from training data through initial features to obtain an initial feature map of a target object; processing the initial feature map through a first initial recognition neural network model toobtain a first prediction result; processing the initial feature map through a second initial recognition neural network model to obtain a second prediction result; procssing the initial feature mapthrough an initial segmentation neural network model to obtain a segmentation result; and updating the initial feature extraction neural network model according to the first prediction result, the second prediction result and the segmentation result to obtain a trained feature extraction neural network model. Through the process, the network capacity, the generalization ability and the extractionprecision of the feature extraction network model can be increased; meanwhile, the segmentation result obtained by using the model can enable the feature extraction network model obtained by trainingto have higher convergence degree and accuracy.

Description

technical field [0001] The present disclosure relates to the field of computer technology, in particular to a neural network training and image recognition method, device, device and storage medium. Background technique [0002] Pedestrian re-identification refers to a query picture of a pedestrian, and it is necessary to find all pictures of the same person in a large-scale data set. With the development of deep learning technology, pedestrian re-identification can be realized by neural network. [0003] However, as the task of pedestrian re-identification becomes more and more complex, the requirements for the recognition accuracy of the neural network are higher, and how to improve the recognition ability of the neural network has become an urgent problem to be solved. Contents of the invention [0004] The present disclosure proposes a training scheme for a neural network. [0005] According to an aspect of the present disclosure, a method for training a neural netwo...

AI Technical Summary

Problems solved by technology

[0003] However, as the task of pedestrian re-identification becomes more and more complex, the requirements for the recognition accuracy of the neural network are higher, how to improve the recognition ability of the neural network has become an urgent problem to be solved

Images