Capsule network multi-feature extraction method based on attention mechanism

Abstract

The invention discloses a capsule network multi-feature recognition and extraction method based on an attention mechanism. The method comprises the following steps: (1) designing an NCap network and constructing an attention mechanism capsule network framework by using the NCap network; (2) inputting an image training set to the attention mechanism capsule network, completing the recognition and extraction of image features after the attention mechanism capsule network is trained and learned, and generating a corresponding optimal training model; (3) inputting a to-be-recognized image to the attention mechanism capsule network, wherein the attention mechanism capsule network loads the optimal network model and recognizes image features; and (4) outputting a recognition result of the to-be-recognized image by the attention mechanism capsule network. The invention provides an attention mechanism-based thought of fusing a convolutional network mechanism and a capsule network structure, the relative position and direction of the image are recorded during training, the parameter quantity is reduced, and the recognition efficiency and accuracy are effectively improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and specifically relates to an image multi-feature recognition and classification method based on an attention mechanism combined with a capsule network in the technical field of image recognition. The invention can be used for extracting and identifying key information features of images. Background technique [0002] In recent years, the technical fields such as target recognition and feature extraction have developed from single-attribute recognition to multi-attribute recognition, and the increasing maturity of its technology has greatly promoted the rapid innovation of re-identification technology. However, there are still some difficulties in the accurate classification of multi-attribute recognition, such as the high dimensionality of pixels, low resolution and noise interference. At present, the methods of target recognition and feature extraction basically use convolutional neu...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem of low accuracy in the multi-attribute feature recognition process and the loss of important information such as the position and direction of the convolutional neural network during the training process, the present invention uses the convolutional neural network as the main method for the current fields of target recognition and feature extraction. Based on the status quo of the current situation, a multi-feature extraction and recognition method based on the attention mechanism of the capsule network is proposed.

Images