Open-vocabulary Instance Segmentation

Open-Vocabulary Instance Segmentation is an instance segmentation framework that extends conventional instance segmentation by enabling the detection and delineation of object instances across a broad, potentially unseen vocabulary of categories at test time, leveraging external knowledge sources such as language models to generalize beyond the fixed set of classes observed during training.

13

Publications

1.2K

Citations

76

Authors

21

Institutions

Publications per year

2017–2026

13

Authors

76

  1. PY
    Pengxiang Yan

    Sun Yat-sen University, China

    2 publications · 84 citations

    2

    h-index

  2. LS
    Lin Song

    XinHua Hospital, China

    2 publications · 589 citations

    2

    h-index

  3. YS
    Ying Shan

    Tencent (China), China

    2 publications · 589 citations

    2

    h-index

  4. XZ
    Xueyan Zou

    University of Wisconsin–Madison, United States

    2 publications · 271 citations

    2

    h-index

  5. XW
    Xinggang Wang

    Huazhong University of Science and Technology, China

    2 publications · 589 citations

    2

    h-index

1–5 of 76

Venues

  1. arXiv (Cornell University)

    Repository · 4 publications · 86 citations

    4

    h-index

  2. 1

    h-index

  3. Pattern Recognition

    1 publications · 18 citations

1–3