Next Article in Journal
Dynamic Analysis of the Median Nerve in Carpal Tunnel Syndrome from Ultrasound Images Using the YOLOv5 Object Detection Model
Previous Article in Journal
Nutritional Factors Associated with Dental Caries across the Lifespan: A Review
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 13
Issue 24
10.3390/app132413255
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation

by
Hyunmin Gwak
1,
Yongho Jeong
2,
Chanyeong Kim
1,
Yonghak Lee
1,
Seongmin Yang
1 and
Sunghwan Kim
1,*
1
Department of Applied Statistics, Konkuk University, Seoul 05029, Republic of Korea
2
AI Analytics Team, Mustree, Seoul 05029, Republic of Korea
*
Author to whom correspondence should be addressed.
Appl. Sci. 2023, 13(24), 13255; https://doi.org/10.3390/app132413255
Submission received: 24 October 2023 / Revised: 15 November 2023 / Accepted: 20 November 2023 / Published: 14 December 2023
Browse Figures
Versions Notes

Abstract

The key to semi-supervised semantic segmentation is to assign the appropriate pseudo-label to the pixels of unlabeled images. Recently, various approaches to consistency-based training and the filtering of reliable pseudo-labels have shown remarkable results. Nonetheless, there are still issues to be addressed. We find that recent approaches have specific problems in common. In pseudo-labels for training unlabeled images, we confirm that false foreground class pseudo-labels are mostly caused by background class confusion, not confusion between different foreground classes. To solve this problem, we propose a foreground and background discrimination model for semi-supervised semantic segmentation. Our proposed model is trained using a novel approach called multi-view integrated ensemble (MVIE) via output perturbation. Experimental results in various partition protocols show that our approach outperforms the existing state of the art (SOTA) in binary prediction on unlabeled data, and the segmentation model trained with the help of our model outperforms existing models.
Keywords: deep learning; image semantic segmentation; semi-supervised semantic segmentation; reliable pseudo-labels deep learning; image semantic segmentation; semi-supervised semantic segmentation; reliable pseudo-labels

Share and Cite

MDPI and ACS Style

Gwak, H.; Jeong, Y.; Kim, C.; Lee, Y.; Yang, S.; Kim, S. A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation. Appl. Sci. 2023, 13, 13255. https://doi.org/10.3390/app132413255

AMA Style

Gwak H, Jeong Y, Kim C, Lee Y, Yang S, Kim S. A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation. Applied Sciences. 2023; 13(24):13255. https://doi.org/10.3390/app132413255

Chicago/Turabian Style

Gwak, Hyunmin, Yongho Jeong, Chanyeong Kim, Yonghak Lee, Seongmin Yang, and Sunghwan Kim. 2023. "A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation" Applied Sciences 13, no. 24: 13255. https://doi.org/10.3390/app132413255

APA Style

Gwak, H., Jeong, Y., Kim, C., Lee, Y., Yang, S., & Kim, S. (2023). A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation. Applied Sciences, 13(24), 13255. https://doi.org/10.3390/app132413255

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop