Attention Guided Enhancement Network for Weakly Supervised Semantic Segmentation  

作　　者：ZHANG Zhe WANG Bilin YU Zhezhou ZHAO Fengzhi 

机构地区：[1]College of Computer Science and Technology,Jilin University,Changchun 130012,China [2]Key Laboratory for Symbol Computation and Knowledge Engineering of National Education Ministry,Changchun 130012,China

出　　处：《Chinese Journal of Electronics》2023年第4期896-907,共12页电子学报（英文版）

基　　金：This work was supported by the Development Project of Jilin Province of China(20200801033GH,2020122328JC);the Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education,Jilin University(20180520017JH);the Fundamental Research Funds for the Central Universities,JLU.

摘　　要：Weakly supervised semantic segmentation using only image-level labels is critical since it alleviates the need for expensive pixel-level labels.Most cuttingedge methods adopt two-step solutions that learn to produce pseudo-ground-truth using only image-level labels and then train off-the-shelf fully supervised semantic segmentation network with these pseudo labels.Although these methods have made significant progress,they also increase the complexity of the model and training.In this paper,we propose a one-step approach for weakly supervised image semantic segmentation—attention guided enhancement network(AGEN),which produces pseudopixel-level labels under the supervision of image-level labels and trains the network to generate segmentation masks in an end-to-end manner.Particularly,we employ class activation maps(CAM)produced by different layers of the classification branch to guide the segmentation branch to learn spatial and semantic information.However,the CAM produced by the lower layer can capture the complete object region but with many noises.Thus,the self-attention module is proposed to enhance object regions adaptively and suppress irrelevant object regions,further boosting the segmentation performance.Experiments on the Pascal VOC 2012 dataset demonstrate that AGEN outperforms alternative state-of-the-art weakly supervised semantic segmentation methods exclusively relying on image-level labels.

关 键 词：Weakly-supervised learning Semantic segmentation Convolutional neural networks Self-attention mechanism END-TO-END 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]