Image background segmentation and recognition method based on convolution neural network

Abstract

The invention discloses an image background segmentation and recognition method based on a convolution neural network. The method uses the convolution neural network to learn a sample image, trains toobtain a background segmentation model and a recognition classification model, and then performs background segmentation and recognition classification of the image according to the model. When the convolution neural network is applied to the image recognition and classification, the invention combines the influence of the image background on the recognition and classification, and the backgroundsegmentation model converts the whole connection layer of the convolution neural network into the convolution layer. The invention improves the effect of image optimization segmentation, and makes the image background segmentation model have wide applicability. Finally, after the image background segmentation is realized by using the convolution neural network model, the image is reused for imagerecognition and classification, so as to improve the accuracy of recognition and classification.

Description

technical field [0001] The invention belongs to image background segmentation and recognition classification, in particular to an image background segmentation and recognition method based on a convolutional neural network. Background technique [0002] There are two main implementations of traditional image segmentation, one is the threshold method, the threshold segmentation method is to simply use one or several thresholds to divide the histogram of the image into several categories, and the gray value in the image is in the same gray category The pixels belonging to the same class can be divided into global threshold segmentation and local threshold segmentation. The simplest form of threshold method can only generate binary images to distinguish two different classes. In addition, it only considers the value of the pixel itself, and generally does not consider the spatial characteristics of the image, so it is very sensitive to noise, and it does not Considering useful...

AI Technical Summary

Problems solved by technology

On the one hand, when convolutional neural networks are currently applied to image recognition and classification, most models do not consider the impact of image background on recognition and classification, and directly use the original image into the model for recognition, which will affect the recognition results

On the other hand, now is the era of information technology and the era of big data. It is difficult for traditional segmentation algorithms to solve the problem of processing such a large amount of data, so it takes a lot of time to process these data to achieve the purpose of image segmentation.

And because the background differences of different types are relatively large, according to the traditional image segmentation method to deal with different types of images, most of the images cannot achieve satisfactory results.

In terms of image recognition, almost no one cares about the impact of the background on image recognition, and this is indeed a problem that cannot be ignored, because some backgrounds seriously affect the effect of its recognition and classification

Images