Image semantic segmentation method and system based on multi-scale feature and foreground and background comparison

Abstract

The invention relates to an image semantic segmentation method and system based on multi-scale feature and foreground and background comparison. The method comprises the following steps of: preprocessing an image, performing feature coding, optimizing shallow features in a coding process, and performing feature decoding by combining the image and the shallow features on the basis of a pixel rearrangement technology and utilizing dense connection to obtain a semantic segmentation probability graph so as to finish the establishment of a core neural network of a semantic segmentation model; then,based on the built core neural network, carrying out data enhancement on the annotation data set, calculating semantic segmentation loss and auxiliary edge detection loss for iteratively updating parameters in the network until convergence is achieved, and completing training of the model; and finally, in combination with the built core neural network and the trained network parameters, selectingthe item with the maximum probability from each position in the obtained semantic segmentation probability graph as the classification of the pixel position to obtain a final semantic segmentation result. The method is beneficial to improving the accuracy and robustness of image semantic segmentation, and the system can be used for customizing an insurance policy system and is used for enhancingand beautifying the image quality of an insurance policy cover image uploaded by a user, filtering sensitive images and the like.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to an image semantic segmentation method and system based on multi-scale features compared with foreground and background. Background technique [0002] Image semantic segmentation is an important branch of computer vision in the field of artificial intelligence and an important part of image understanding in machine vision. Image semantic segmentation is to accurately classify each pixel in the image to its category, making it consistent with the visual representation of the image itself, so the image semantic segmentation task is also called pixel-level image classification task. [0003] Due to the similarity between image semantic segmentation and image classification, various image classification networks are often used as the backbone of the image semantic segmentation network after removing the last fully connected layer, and are interchangeable with each other. Some...

AI Technical Summary

Problems solved by technology

However, the former has a low information utilization rate due to the large-scale expansion convolution, and the latter adopts pooling to cause the loss of spatial information of the features, and both fail to take into account the correlation of the foreground and background.

In the existing semantic segmentation methods, in the decoding process, the features are generally expanded step by step by using methods such as transposed convolution or bilinear interpolation, so the feature size is increased step by step, and it is impossible to effectively reuse the reconstructed features.

And in this process, shallow features are often added to optimize the decoding process, but there is no clear optimization target for shallow features, so the ability to repair details in the reconstruction process is slightly insufficient

Images