{
  "id": "2112.10982",
  "version": "v2",
  "published": "2021-12-21T04:44:57.000Z",
  "updated": "2022-03-22T19:17:04.000Z",
  "title": "Generalized Few-Shot Semantic Segmentation: All You Need is Fine-Tuning",
  "authors": [
    "Josh Myers-Dean",
    "Yinan Zhao",
    "Brian Price",
    "Scott Cohen",
    "Danna Gurari"
  ],
  "comment": "Includes supplementary materials",
  "categories": [
    "cs.CV",
    "cs.LG"
  ],
  "abstract": "Generalized few-shot semantic segmentation was introduced to move beyond only evaluating few-shot segmentation models on novel classes to include testing their ability to remember base classes. While the current state-of-the-art approach is based on meta-learning, it performs poorly and saturates in learning after observing only a few shots. We propose the first fine-tuning solution, and demonstrate that it addresses the saturation problem while achieving state-of-the-art results on two datasets, PASCAL-5i and COCO-20i. We also show that it outperforms existing methods, whether fine-tuning multiple final layers or only the final layer. Finally, we present a triplet loss regularization that shows how to redistribute the balance of performance between novel and base categories so that there is a smaller gap between them.",
  "revisions": [
    {
      "version": "v2",
      "updated": "2022-03-22T19:17:04.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "generalized few-shot semantic segmentation",
      "evaluating few-shot segmentation models",
      "current state-of-the-art approach",
      "triplet loss regularization",
      "fine-tuning multiple final layers"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}