{
  "id": "2006.10990",
  "version": "v1",
  "published": "2020-06-19T07:35:25.000Z",
  "updated": "2020-06-19T07:35:25.000Z",
  "title": "Cross-denoising Network against Corrupted Labels in Medical Image Segmentation with Domain Shift",
  "authors": [
    "Qinming Zhang",
    "Luyan Liu",
    "Kai Ma",
    "Cheng Zhuo",
    "Yefeng Zheng"
  ],
  "comment": "Accepted by IJCAI 2020",
  "categories": [
    "cs.CV",
    "cs.LG",
    "eess.IV"
  ],
  "abstract": "Deep convolutional neural networks (DCNNs) have contributed many breakthroughs in segmentation tasks, especially in the field of medical imaging. However, \\textit{domain shift} and \\textit{corrupted annotations}, which are two common problems in medical imaging, dramatically degrade the performance of DCNNs in practice. In this paper, we propose a novel robust cross-denoising framework using two peer networks to address domain shift and corrupted label problems with a peer-review strategy. Specifically, each network performs as a mentor, mutually supervised to learn from reliable samples selected by the peer network to combat with corrupted labels. In addition, a noise-tolerant loss is proposed to encourage the network to capture the key location and filter the discrepancy under various noise-contaminant labels. To further reduce the accumulated error, we introduce a class-imbalanced cross learning using most confident predictions at the class-level. Experimental results on REFUGE and Drishti-GS datasets for optic disc (OD) and optic cup (OC) segmentation demonstrate the superior performance of our proposed approach to the state-of-the-art methods.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-06-19T07:35:25.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "medical image segmentation",
      "corrupted label",
      "cross-denoising network",
      "deep convolutional neural networks",
      "peer network"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}