Semantic segmentation method and system based on edge dense reconstruction for streetscape understanding

Abstract

The invention relates to a semantic segmentation method and system based on edge dense reconstruction for streetscape understanding, and the method comprises the steps: carrying out the preprocessingof an input image of a training set, enabling the image to be standardized, and obtaining preprocessed images with the same size; extracting general features by using a convolutional network, then obtaining three-level context space pyramid fusion features, and extracting coding features by using the two parts of cascade connection as a coding network; acquiring semi-input size encoding features by using the encoding features, acquiring edge features based on a convolutional network, and reconstructing image resolution by taking a dense network fused with the edge features as a decoding network in combination with the semi-input size encoding features, and acquiring decoding features; calculating semantic segmentation loss and auxiliary supervision edge loss, and training the deep neural network by taking minimization of weighted sum loss of the semantic segmentation loss and the auxiliary supervision edge loss as a target; and performing semantic segmentation on the to-be-segmented image by using the deep neural network model, and outputting a segmentation result. The method and the system are beneficial to improving the accuracy and robustness of image semantic segmentation.

Description

Technical field [0001] The present invention relates to the technical field of computer vision, in particular to a semantic segmentation method and system based on dense edge reconstruction for street scene understanding. 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 into its own category to make it consistent with the visual representation of the image itself. Therefore, the image semantic segmentation task is also called a 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 can be replaced with each other. S...

AI Technical Summary

Problems solved by technology

At the same time, in the encoding process, in order to better capture the characteristics of objects of different sizes, different receptive fields and scale information are often combined, such as the spatial pyramid pooling technology with holes, but this technology expands the interval of the convolution kernel and ignores the internal pixels. At the same time, it fails to combine more global context information to make up for the lack of self-expression ability

At the same time, in the existing semantic segmentation methods, the resolution is often simply restored based on the previous level of features in the decoding process, and then combined with the shallow features of the corresponding size to compensate for the information loss in the encoding process, which has not effectively The reuse of effective features in the resolution reconstruction process also fails to solve the problem of blurred object boundaries after image resolution reconstruction

Images