Sökning: id:"swepub:oai:research.chalmers.se:97ad91cb-afe3-40f2-9074-bcc5c0ce5776" >
Rethinking Long-Tai...
Rethinking Long-Tailed Visual Recognition with Dynamic Probability Smoothing and Frequency Weighted Focusing
-
Nah, Wan Jun (författare)
-
Ng, Chun Chet (författare)
-
- Lin, Che-Tsung, 1979 (författare)
- Chalmers tekniska högskola,Chalmers University of Technology
-
visa fler...
-
Lee, Yeong Khang (författare)
-
Kew, Jie Long (författare)
-
Tan, Zhi Qin (författare)
-
Chan, Chee Seng (författare)
-
- Zach, Christopher, 1974 (författare)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Lai, Shang Hong (författare)
- National Tsing Hua University
-
visa färre...
-
(creator_code:org_t)
- 2023
- 2023
- Engelska.
-
Ingår i: Proceedings - International Conference on Image Processing, ICIP. - 1522-4880. ; , s. 435-439
- Relaterad länk:
-
https://doi.org/10.1...
-
visa fler...
-
https://research.cha...
-
visa färre...
Abstract
Ämnesord
Stäng
- Deep learning models trained on long-tailed (LT) datasets often exhibit bias towards head classes with high frequency. This paper highlights the limitations of existing solutions that combine class- and instance-level re-weighting loss in a naive manner. Specifically, we demonstrate that such solutions result in overfitting the training set, significantly impacting the rare classes. To address this issue, we propose a novel loss function that dynamically reduces the influence of outliers and assigns class-dependent focusing parameters. We also introduce a new long-tailed dataset, ICText-LT, featuring various image qualities and greater realism than artificially sampled datasets. Our method has proven effective, outperforming existing methods through superior quantitative results on CIFAR-LT, Tiny ImageNet-LT, and our new ICText-LT datasets. The source code and new dataset are available at \url{https://github.com/nwjun/FFDS-Loss}
Ämnesord
- TEKNIK OCH TEKNOLOGIER -- Elektroteknik och elektronik (hsv//swe)
- ENGINEERING AND TECHNOLOGY -- Electrical Engineering, Electronic Engineering, Information Engineering (hsv//eng)
Nyckelord
- Long-tailed Classification
- Weighted-loss
Publikations- och innehållstyp
- kon (ämneskategori)
- ref (ämneskategori)
Hitta via bibliotek
Till lärosätets databas
- Av författaren/redakt...
-
Nah, Wan Jun
-
Ng, Chun Chet
-
Lin, Che-Tsung, ...
-
Lee, Yeong Khang
-
Kew, Jie Long
-
Tan, Zhi Qin
-
visa fler...
-
Chan, Chee Seng
-
Zach, Christophe ...
-
Lai, Shang Hong
-
visa färre...
- Om ämnet
-
- TEKNIK OCH TEKNOLOGIER
-
TEKNIK OCH TEKNO ...
-
och Elektroteknik oc ...
- Artiklar i publikationen
-
Proceedings - In ...
- Av lärosätet
-
Chalmers tekniska högskola