Deeplabv3+ semantic segmentation model based on feature cross attention mechanism

Abstract

Aiming at the problem that the deeplabv3+ model is not accurate in segmentation of the image target edge, the image feature fitting is slow, and the attention information cannot be effectively used. It is proposed to add a feature cross attention module (FCA) to the model. The cross-attention network is composed of two branches and a feature cross attention module. Among them, the shallow branch is used to extract low-level spatial information, and the deep branch is used to extract high-level context features to make important feature extraction more refined. This paper designs and realizes the connection between Feature Cross Attention module and Deeplabv3+ coding module, input the output features of the Deeplabv3+ encoding module into the feature cross attention module for convolution operation to realize the recalibration of the original features. The decoding module of Deeplabv3+ obtains spatial features and channel features from two branches respectively, and then merges the obtained features to obtain more important features. The improved model was validated by the Pascal Voc2012 data set, and the results showed that the ratio of average intersection and the average pixel accuracy were increased by 1.96% and 2.84%, respectively. The model added with FCA can effectively improve the shortcomings of the original model, can segment the target more finely, and better solve the problem of rough segmentation boundary.

Comparison of Different Transfer Learning Methods for Classification of Mangrove Communities Using MCCUNet and UAV Multispectral Images

“…So, there is more room for improving the algorithm. In future research, the classification performance of the algorithm may be further improved by adding further attention mechanisms to the network [45,46] or by replacing the basic structural units of the backbone network [47] to better apply to the identification of mangrove communities.…”

mid-DeepLabv3+: A Novel Approach for Image Semantic Segmentation Applied to African Food Dietary Assessments

“…Attention mechanisms have proven to be effective in computer vision, allowing models to focus on relevant parts of the input by assigning different weights to different regions [ 46 ]. Several works have demonstrated that incorporating attention can enhance the performance of semantic segmentation models, including DeepLabv3+ [ 47 , 48 , 49 ]. In our work, we have chosen to use the SimAM (Simple Attention Module) [ 10 ].…”

Stripe pooling and densely connected DeepLabv3 plus efficient semantic segmentation

“…The overall structure of the improved DeepLabv3 plus [8,9,10] model is shown in Figure 1 below. In the encoder section, MobileneV2 is used as the main part to accelerate prediction speed.…”

“…We propose an improved algorithm based on DeepLabV3 plus, which can effectively optimize the original network structure. By reconstructing the ASPP [12] module in DeepLabv3, utilizing stripe pooling instead of global average pooling, and modifying its void ratio to fuse information from different scales, a denser feature scale range is obtained in a densely connected manner, thereby improving the segmentation accuracy of network recognition.…”