Boundary-Aware Bias Loss for Transformer-Based Aerial Image Segmentation Model
Authors:
- 23 May 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3528-3532
Abstract
Inspired by the tremendous success of the transformer-based model in natural language processing (NLP), many efforts introduce the transformer-based model into the image processing tasks. However, naive transformer models have to down-sample the image resolution to satisfy computational restrictions, thus discarding the local information, which is catastrophic for high-performance remote sensing image segmentation. Hence, this paper proposes a novel trainable boundary-aware bias loss function to enhance transformer-based models of extracting local information. On the Challenging ISPRS Potsdam dataset, two representative transformer-based models achieve remarkable performance improvements, proving the effectiveness of the proposed method.Keywords
This publication has 9 references indexed in Scilit:
- Swin Transformer: Hierarchical Vision Transformer using Shifted Windows2021
- Conformer: Local Features Coupling Global Representations for Visual Recognition2021
- SEMEDA: Enhancing segmentation precision with semantic edge aware lossPattern Recognition, 2020
- Web-Net: A Novel Nest Networks with Ultra-Hierarchical Sampling for Building Extraction from Aerial ImageriesRemote Sensing, 2019
- Simultaneous extraction of roads and buildings in remote sensing imagery with convolutional neural networksISPRS Journal of Photogrammetry and Remote Sensing, 2017
- DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- Semantic Segmentation of Small Objects and Modeling of Uncertainty in Urban Remote Sensing Images Using Deep Convolutional Neural Networks2016
- Fully convolutional networks for semantic segmentation2015
- ImageNet: A large-scale hierarchical image database2009