{
  "id": "2205.03776",
  "version": "v1",
  "published": "2022-05-08T04:00:28.000Z",
  "updated": "2022-05-08T04:00:28.000Z",
  "title": "SparseTT: Visual Tracking with Sparse Transformers",
  "authors": [
    "Zhihong Fu",
    "Zehua Fu",
    "Qingjie Liu",
    "Wenrui Cai",
    "Yunhong Wang"
  ],
  "comment": "Accepted by IJCAI2022 as a long oral presentation",
  "categories": [
    "cs.CV"
  ],
  "abstract": "Transformers have been successfully applied to the visual tracking task and significantly promote tracking performance. The self-attention mechanism designed to model long-range dependencies is the key to the success of Transformers. However, self-attention lacks focusing on the most relevant information in the search regions, making it easy to be distracted by background. In this paper, we relieve this issue with a sparse attention mechanism by focusing the most relevant information in the search regions, which enables a much accurate tracking. Furthermore, we introduce a double-head predictor to boost the accuracy of foreground-background classification and regression of target bounding boxes, which further improve the tracking performance. Extensive experiments show that, without bells and whistles, our method significantly outperforms the state-of-the-art approaches on LaSOT, GOT-10k, TrackingNet, and UAV123, while running at 40 FPS. Notably, the training time of our method is reduced by 75% compared to that of TransT. The source code and models are available at https://github.com/fzh0917/SparseTT.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2022-05-08T04:00:28.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "visual tracking",
      "sparse transformers",
      "search regions",
      "relevant information",
      "sparse attention mechanism"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}