Learning Task-Specific Generalized Convolutions in the Permutohedral Lattice

被引:1
|
作者
Wannenwetsch, Anne S. [1 ]
Kiefel, Martin [2 ]
Gehler, Peter V. [2 ]
Roth, Stefan [1 ]
机构
[1] Tech Univ Darmstadt, Darmstadt, Germany
[2] Amazon, Tubingen, Germany
来源
关键词
D O I
10.1007/978-3-030-33676-9_24
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Dense prediction tasks typically employ encoder-decoder architectures, but the prevalent convolutions in the decoder are not image-adaptive and can lead to boundary artifacts. Different generalized convolution operations have been introduced to counteract this. We go beyond these by leveraging guidance data to redefine their inherent notion of proximity. Our proposed network layer builds on the permutohedral lattice, which performs sparse convolutions in a high-dimensional space allowing for powerful non-local operations despite small filters. Multiple features with different characteristics span this permutohedral space. In contrast to prior work, we learn these features in a task-specific manner by generalizing the basic permutohedral operations to learnt feature representations. As the resulting objective is complex, a carefully designed framework and learning procedure are introduced, yielding rich feature embeddings in practice. We demonstrate the general applicability of our approach in different joint upsampling tasks. When adding our network layer to state-of-the-art networks for optical flow and semantic segmentation, boundary artifacts are removed and the accuracy is improved.
引用
收藏
页码:345 / 359
页数:15
相关论文
共 50 条
  • [41] Learning Task-specific Activation Functions using Genetic Programming
    Basirat, Mina
    Roth, Peter M.
    PROCEEDINGS OF THE 14TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS (VISAPP), VOL 5, 2019, : 533 - 540
  • [42] Learning an EMG Controlled Game: Task-Specific Adaptations and Transfer
    van Dijk, Ludger
    van der Sluis, Corry K.
    van Dijk, Hylke W.
    Bongers, Raoul M.
    PLOS ONE, 2016, 11 (08):
  • [43] Learning task-specific discriminative representations for multiple object tracking
    Han Wu
    Jiahao Nie
    Ziming Zhu
    Zhiwei He
    Mingyu Gao
    Neural Computing and Applications, 2023, 35 : 7761 - 7777
  • [44] Task-specific experience and task-specific talent: Decomposing the productivity of high school teachers
    Cook, Jason B.
    Mansfield, Richard K.
    JOURNAL OF PUBLIC ECONOMICS, 2016, 140 : 51 - 72
  • [45] Task-specific visualization design
    Treinish, Lloyd A.
    IEEE Computer Graphics and Applications, 19 (05): : 72 - 77
  • [46] Task-Specific Drinking Tremor
    Benedek, Krisztina
    Biernat, Heidi Bryde
    Thomsen, Carsten Eckhart
    Bakke, Merete
    JOURNAL OF MOVEMENT DISORDERS, 2023, 16 (01) : 98 - 100
  • [47] Task-specific visualization design
    Treinish, LA
    IEEE COMPUTER GRAPHICS AND APPLICATIONS, 1999, 19 (05) : 72 - 77
  • [48] Task-specific craniocervical dystonia
    Schramm, Axel
    Naumann, Markus
    Reiners, Karlheinz
    Classen, Joseph
    MOVEMENT DISORDERS, 2008, 23 (07) : 1041 - 1043
  • [49] Task-specific Dystonias A Review
    Torres-Russotto, Diego
    Perlmutter, Joel S.
    YEAR IN NEUROLOGY 2008, 2008, 1142 : 179 - 199
  • [50] Task-specific ionic liquids
    Davis, JH
    CHEMISTRY LETTERS, 2004, 33 (09) : 1072 - 1077