A multi-label classification method and system

Abstract

The embodiment of the invention provides a multi-label classification method and system. The method comprises: obtaining all target objects in the image to be tested according to the image to be tested and the trained improved neural network, wherein, the improved neural network is obtained by combining the neural network and the attention mechanism. The embodiment of the invention provides a multi-label classification method, which combines an attention mechanism with a neural network to highlight the importance of each target object in an image to be measured, so that each target object canbe more accurately identified when the multi-target is extracted, and the problem of missing identification in the prior art is avoided.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of object recognition and classification, and in particular to a multi-label classification method and system. Background technique [0002] In the process of multi-label classification, a picture often contains multiple targets. In the prior art, for image multi-label classification tasks, the main deep learning method is to determine an input picture size and then train on the data set. , by setting multiple binary classifiers, if the output of the binary classifier of a certain class is 0, it means that the picture contains this class. [0003] However, there are many cases of false detection by this method. If the response value of the relevant response area on the last feature layer is small, the model will not be able to determine whether this category is included. Contents of the invention [0004] Embodiments of the present invention provide a method and system for classifyin...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a method and system for classifying labels to solve the problem that multiple targets cannot be accurately identified in the prior art

Images