Computer Science > Computer Vision and Pattern Recognition
[Submitted on 18 Jun 2024]
Title:Unleashing the Potential of Open-set Noisy Samples Against Label Noise for Medical Image Classification
View PDF HTML (experimental)Abstract:The challenge of addressing mixed closed-set and open-set label noise in medical image classification remains largely unexplored. Unlike natural image classification where there is a common practice of segregation and separate processing of closed-set and open-set noisy samples from clean ones, medical image classification faces difficulties due to high inter-class similarity which complicates the identification of open-set noisy samples. Moreover, prevailing methods do not leverage the full potential of open-set noisy samples for label noise mitigation, often leading to their exclusion or application of uniform soft labels. To address these issues, we propose an Extended Noise-robust Contrastive and Open-set Feature Augmentation (ENCOFA) framework. ENCOFA includes the Extended Noise-robust Supervised Contrastive (ENSC) Loss, which aids in distinguishing features across classes. The ENSC loss regards open-set noisy samples as an extended class and mitigates label noise by weighting contrastive pairs with label reliability. Furthermore, we develop an Open-set Feature Augmentation (OSFeatAug) module that enriches the features of open-set samples, utilizing the model's extra capacity to prevent overfitting to noisy data. We conducted experiments on a synthetic noisy dataset and a real-world noisy dataset. Our results indicate the superiority of ENCOFA and the effectiveness of leveraging the open-set noisy samples to combat label noise.
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.