Target detection method and device and storage medium thereof

Abstract

The invention discloses a target detection method and device and a storage medium thereof. The method comprises the steps of receiving a marked model image and an unmarked model image, and training the marked model image and the unannotated model image through a pre-built student network and a pre-built teacher network according to an average teacher strategy to obtain marked model image loss andunannotated model image loss; obtaining a training model according to the marked model image loss and the unannotated model image loss; receiving a to-be-identified image; and detecting the to-be-identified image by using the training model. According to the invention, the marked model image and the unmarked model image are trained through the neural network according to the average teacher strategy to obtain the training model, and the training model is utilized to detect the images, so that the accuracy of recognition and detection can be greatly improved.

Description

technical field [0001] The present invention relates to the field of computers, in particular to a target detection method, device and storage medium thereof. Background technique [0002] In recent years, vision research on generating predictive labels from synthetic data based on deep learning models has attracted increasing attention. For example, in some images, it is necessary to identify the target objects such as cars and people inside, and use tags to locate and label the target objects in an appropriate range. Although the deep neural network has been proved to be an efficient method for learning visual models under large data sets, errors are still common when simply using the learning model obtained by the deep neural network for image synthesis to generate predicted labels. [0003] Therefore, the inventors found that there are at least the following problems in the prior art. When simply using the learning model obtained by the deep neural network to generate a...

AI Technical Summary

Problems solved by technology

Although the deep neural network has been proved to be an efficient method for learning visual models under large data sets, there are still widespread errors when simply using the learning model obtained by the deep neural network for image synthesis to generate predicted labels

[0003] Therefore, the inventors found that there are at least the following problems in the prior art. When simply using the learning model obtained by the deep neural network to generate a prediction label to identify the recognition object, there are still common technical problems of errors

Images