PlaneRecTR: Unified Query Learning for 3D Plane Recovery from a Single View

被引:1
作者
Shi, Jingjia [1 ]
Zhi, Shuaifeng [1 ]
Xu, Kai [1 ]
机构
[1] Natl Univ Def Technol, Changsha, Hunan, Peoples R China
来源
2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2023) | 2023年
关键词
D O I
10.1109/ICCV51070.2023.00860
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
3D plane recovery from a single image can usually be divided into several subtasks of plane detection, segmentation, parameter estimation and possibly depth estimation. Previous works tend to solve it by either extending the RCNN-based segmentation network or the dense pixel embedding-based clustering framework. However, none of them tried to integrate above related subtasks into a unified framework but treated them separately and sequentially, which we suspect is potentially a main source of performance limitation for existing approaches. Motivated by this finding and the success of query-based learning in enriching reasoning among semantic entities, in this paper, we propose PlaneRecTR, a Transformer- based architecture, which for the first time unifies all subtasks related to single-view plane recovery with a single compact model. Extensive quantitative and qualitative experiments demonstrate that our proposed unified learning achieves mutual benefits across subtasks, obtaining a new state-of-the-art performance on public ScanNet and NYUv2-Plane datasets.
引用
收藏
页码:9343 / 9352
页数:10
相关论文
共 39 条
  • [1] [Anonymous], 2018, ARXIV180311288
  • [2] Contour Detection and Hierarchical Image Segmentation
    Arbelaez, Pablo
    Maire, Michael
    Fowlkes, Charless
    Malik, Jitendra
    [J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2011, 33 (05) : 898 - 916
  • [3] Barinova Olga, 2008, P EUR C COMP VIS ECC
  • [4] Carion Nicolas, 2020, Computer Vision - ECCV 2020. 16th European Conference. Proceedings. Lecture Notes in Computer Science (LNCS 12346), P213, DOI 10.1007/978-3-030-58452-8_13
  • [5] Cheng B, 2021, ADV NEUR IN, V34
  • [6] Cheng Bowen, 2022, P IEEE C COMP VIS PA
  • [7] Dai Angela, 2017, INT C COMP VIS PATT, V2, P3
  • [8] Delage Erick, 2007, P INT S ROB RES ISRR
  • [9] Dosovitskiy Alexey, 2020, INT C LEARN REPR ICL, DOI DOI 10.48550/ARXIV.2010.11929
  • [10] Fouhey David Ford, 2014, P EUR C COMP VIS ECC, P2