Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Co-Seg: An Image Segmentation Framework Against Label Corruption

Jan 31, 2021
Ziyi Huang, Haofeng Zhang, Andrew Laine, Elsa Angelini, Christine Hendon, Yu Gan

Share this with someone who'll enjoy it:

Supervised deep learning performance is heavily tied to the availability of high-quality labels for training. Neural networks can gradually overfit corrupted labels if directly trained on noisy datasets, leading to severe performance degradation at test time. In this paper, we propose a novel deep learning framework, namely Co-Seg, to collaboratively train segmentation networks on datasets which include low-quality noisy labels. Our approach first trains two networks simultaneously to sift through all samples and obtain a subset with reliable labels. Then, an efficient yet easily-implemented label correction strategy is applied to enrich the reliable subset. Finally, using the updated dataset, we retrain the segmentation network to finalize its parameters. Experiments in two noisy labels scenarios demonstrate that our proposed model can achieve results comparable to those obtained from supervised learning trained on the noise-free labels. In addition, our framework can be easily implemented in any segmentation algorithm to increase its robustness to noisy labels.

   Access Paper Source

Share this with someone who'll enjoy it: