Distribution-aware margin calibration for semantic segmentation in images

DOI10.1007/S11263-021-01533-0zbMATH Open1489.68329arXiv2112.11554OpenAlexW3211354896MaRDI QIDQ2117008FDOQ2117008

Min Xu, Jian Zhang, Jiebo Luo, Zhibin Li, Litao Yu, Yongsheng Gao

Publication date: 18 March 2022

Published in: International Journal of Computer Vision (Search for Journal in Brave)

Abstract: The Jaccard index, also known as Intersection-over-Union (IoU), is one of the most critical evaluation metrics in image semantic segmentation. However, direct optimization of IoU score is very difficult because the learning objective is neither differentiable nor decomposable. Although some algorithms have been proposed to optimize its surrogates, there is no guarantee provided for the generalization ability. In this paper, we propose a margin calibration method, which can be directly used as a learning objective, for an improved generalization of IoU over the data-distribution, underpinned by a rigid lower bound. This scheme theoretically ensures a better segmentation performance in terms of IoU score. We evaluated the effectiveness of the proposed margin calibration method on seven image datasets, showing substantial improvements in IoU score over other learning objectives using deep segmentation models.

Full work available at URL: https://arxiv.org/abs/2112.11554

Recommendations

• The role of optimum connectivity in image segmentation: can the algorithm learn object information during the process?
• Subset selection for visualization of relevant image fractions for deep learning based semantic image segmentation
• scientific article; zbMATH DE number 7109071
• Reaction-diffusion model applied to enhancing U-net accuracy for semantic image segmentation
• Revisiting Deep Structured Models for Pixel-Level Labeling with Gradient-Based Inference

zbMATH Keywords

Jaccard indexsemantic segmentationIoUmargin calibration

Mathematics Subject Classification ID

Machine vision and scene understanding (68T45)

Cites Work

• Title not available (Why is that?)

Uses Software

• DeepLab
• U-Net
• PASCAL VOC
• Cityscapes
• Mapillary Vistas
• ADE20k
• BDD100k

This page was built for publication: Distribution-aware margin calibration for semantic segmentation in images

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2117008)