Real-time image semantic segmentation method based on lightweight full convolutional neural network

Abstract

The invention discloses a real-time image semantic segmentation method based on a lightweight full convolutional neural network. The method comprises the following steps of 1) constructing a full convolutional neural network by using the design elements of a lightweight neural network, wherein the network totally comprises three stages of a feature extension stage, a feature processing stage and acomprehensive prediction stage, and the feature processing stage uses a multi-receptive field feature fusion structure, a multi-size convolutional fusion structure and a receptive field amplificationstructure; 2) at a training stage, training the network by using a semantic segmentation data set, using a cross entropy function as a loss function, using an Adam algorithm as a parameter optimization algorithm, and using an online difficult sample retraining strategy in the process; and 3) at a test stage, inputting the test image into the network to obtain a semantic segmentation result. According to the present invention, the high-precision real-time semantic segmentation method suitable for running on a mobile terminal platform is obtained by adjusting a network structure and adapting asemantic segmentation task while controlling the scale of the model.

Description

Technical field [0001] The invention belongs to the technical field of computer software, relates to image semantic segmentation technology, and specifically is a real-time image semantic segmentation method based on a lightweight full convolutional neural network. Background technique [0002] Image semantic segmentation is an intensive predictive classification task. It needs to predict the classification label of each pixel of the input image. It is often used as a pilot process for scene recognition and automatic obstacle avoidance tasks. It is a hot research topic in the field of computer vision. Since AlexNet has shined in the ImageNet competition in 2012, deep learning has been widely used in the field of computer vision. At present, methods based on deep learning also occupy half of the field of semantic segmentation, most of which use fully convolutional neural networks, and gradually form a common structure that encodes and then decodes. In the encoding stage, depth fe...

AI Technical Summary

Problems solved by technology

[0004] The problem to be solved by the present invention is: the existing semantic segmentation method is difficult to run on the mobile terminal platform under the dual requirements of hardware computing power and real-time performance, or because of hardware constraints, it is difficult to run on the mobile terminal platform, or because of method defects , when running on the mobile platform, the classification accuracy is poor

Images