Image semantic segmentation based on global and local features of deep learning

Abstract

The invention relates to an image semantic segmentation method based on global features and local features of deep learning. The method comprises the following steps: at an encoding end, the basic deep features of an image are extracted by using a convolution neural network model based on deep learning; meanwhile, the features are divided into low-level features and high-level features according to the deep of the convolution layer; the feature fusion module fuses the low-level feature and the high-level feature into the enhanced deep feature; the feature fusion module fuses the low-level feature and the high-level feature into the enhanced deep feature; after the deep feature is acquired, the deep feature is inputted to the decoding end; the cross-entropy loss function is used to train the network and mIoU is used to evaluate the performance of the network. The invention is reasonable in design, and uses the deep convolution neural network model to extract the global and local features of an image, fully utilizes the complementarity of the global features and the local features, further improves the performance by utilizing the stack pooling layer, and effectively improves the accuracy of image semantic segmentation.

Description

technical field [0001] The invention belongs to the technical field of computer vision image semantic segmentation, in particular to an image semantic segmentation method based on global features and local features of deep learning. Background technique [0002] Image semantic segmentation refers to dividing each pixel in the image into different semantic categories by a certain method, realizing the reasoning process from the bottom layer to the high-level semantics, and finally obtaining a pixel-by-pixel semantically labeled segmentation map showing different segmentation regions. Image semantic segmentation is widely used in many computer vision tasks such as street scene recognition and target detection in automotive autonomous driving, drone landing point detection, scene understanding, robot vision, etc. From the machine learning method based on computer vision to the current method based on deep learning, the research of image semantic segmentation algorithm has made ...

AI Technical Summary

Problems solved by technology

These methods rely too much on manually annotated feature libraries, and it is difficult to widely represent image features, which has great limitations in practical applications.

Images