On Power Jaccard Losses for Semantic Segmentation

被引：20

作者：

Duque-Arias, David ^{[1
]}

Velasco-Forero, Santiago ^{[1
]}

Deschaud, Jean-Emmanuel ^{[1
]}

Goulette, Francois ^{[1
]}

Serna, Andres ^{[2
]}

Decenciere, Etienne ^{[1
]}

Marcotegui, Beatriz ^{[1
]}

机构：

[1] PSL Res Univ, MINES ParisTech, Paris, France

[2] Terra3D Res, Paris, France

来源：

VISAPP: PROCEEDINGS OF THE 16TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS - VOL. 5: VISAPP | 2021年

关键词：

Loss Functions; Image Segmentation; Jaccard Loss; Deep Learning; U-Net Architecture;

D O I：

10.5220/0010304005610568

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this work, a new generalized loss function is proposed called power Jaccard to perform semantic segmentation tasks. It is compared with classical loss functions in different scenarios, including gray level and color image segmentation, as well as 3D point cloud segmentation. The results show improved performance, stability and convergence. We made available the code with our proposal with a demonstrative example.

引用

页码：561 / 568

页数：8

共 50 条

[1] Jaccard Metric Losses: Optimizing the Jaccard Index with Soft Labels
Wang, Zifu
Ning, Xuefei
Blaschko, Matthew B.
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
[2] SAL:Selection and Attention Losses for Weakly Supervised Semantic Segmentation
Zhou, Lei
Gong, Chen
Liu, Zhi
Fu, Keren
IEEE TRANSACTIONS ON MULTIMEDIA, 2021, 23 : 1035 - 1048
[3] Jaccard Index in Ensemble Image Segmentation: An Approach
Ogwok, Daniel
Ehlers, Elizabeth Marie
ACM International Conference Proceeding Series, 2022, : 9 - 14
[4] Jaccard Median for Ego-network Segmentation
Zhong, Haodi
Loukides, Grigorios
Conte, Alessio
Pissis, Solon P.
2022 IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM), 2022, : 1353 - 1358
[5] JACCARD INDEX COMPENSATION FOR OBJECT SEGMENTATION EVALUATION
Shi, Ran
Ngan, King Ngi
Li, Songnan
2014 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2014, : 4457 - 4461
[6] Class Enhancement Losses With Pseudo Labels for Open-Vocabulary Semantic Segmentation
Dao, Son Duy
Shi, Hengcan
Phung, Dinh
Cai, Jianfei
IEEE TRANSACTIONS ON MULTIMEDIA, 2024, 26 : 8442 - 8453
[7] Short Text Similarity Calculation Based on Jaccard and Semantic Mixture
Wu, Shushu
Liu, Fang
Zhang, Kai
Communications in Computer and Information Science, 2021, 1363 CCIS : 37 - 45
[8] Ego-Network Segmentation via (Weighted) Jaccard Median
Zhong, Haodi
Loukides, Grigorios
Conte, Alessio
Pissis, Solon P.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2024, 36 (09) : 4646 - 4663
[9] Generation of Smoke Dataset for Power Equipment and Study of Image Semantic Segmentation
Chang, Rong
Mao, Zhengxiong
Hu, Jian
Bai, Haicheng
Pan, Anning
Yang, Yang
Gao, Shan
JOURNAL OF ELECTRICAL AND COMPUTER ENGINEERING, 2024, 2024
[10] Low-power neural networks for semantic segmentation of satellite images
Bahl, Gaetan
Daniel, Lionel
Moretti, Matthieu
Lafarge, Florent
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW), 2019, : 2469 - 2476

← 1 2 3 4 5 →