Self-supervised Difference Detection for Refinement CRF and Seed Interpolation

Wataru Shimoda, Keiji Yanai; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2019, pp. 73-77

Abstract


To minimize annotation costs associated with training of semantic segmentation models, weakly-supervised segmentation approaches have been extensively studied. In this paper, we propose a novel method: Self-Supervised Difference Detection (SSDD) module which evaluates confidence of each of the pixels of segmentation masks and integrate highly confident pixels of two candidate masks.

Related Material


[pdf]
[bibtex]
@InProceedings{Shimoda_2019_CVPR_Workshops,
author = {Shimoda, Wataru and Yanai, Keiji},
title = {Self-supervised Difference Detection for Refinement CRF and Seed Interpolation},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2019}
}