Points to patches: Enabling the use of self-attention for 3D shape recognition

Axel Berg, Magnus Oskarsson, Mark O'Connor

Forskningsoutput: Kapitel i bok/rapport/Conference proceedingKonferenspaper i proceedingPeer review


While the Transformer architecture has become ubiquitous in the machine learning field, its adaptation to 3D shape recognition is non-trivial. Due to its quadratic computational complexity, the self-attention operator quickly becomes inefficient as the set of input points grows larger. Furthermore, we find that the attention mechanism struggles to find useful connections between individual points on a global scale. In order to alleviate these problems, we propose a two-stage Point Transformer-in-Transformer (Point-TnT) approach which combines local and global attention mechanisms, enabling both individual points and patches of points to attend to each other effectively. Experiments on shape classification show that such an approach provides more useful features for downstream tasks than the baseline Transformer, while also being more computationally efficient. In addition, we also extend our method to feature matching for scene reconstruction, showing that it can be used in conjunction with existing scene reconstruction pipelines.
Titel på värdpublikation2022 26th International Conference on Pattern Recognition (ICPR)
FörlagIEEE - Institute of Electrical and Electronics Engineers Inc.
Antal sidor7
ISBN (elektroniskt)978-1-6654-9062-7
ISBN (tryckt)978-1-6654-9062-7
StatusPublished - 2022 aug. 21
Evenemang26TH International Conference on Pattern Recognition, 2022 - Montreal, Kanada
Varaktighet: 2022 aug. 212022 aug. 25


NamnInternational Conference on Pattern Recognition
ISSN (tryckt)1051-4651
ISSN (elektroniskt)2831-7475


Konferens26TH International Conference on Pattern Recognition, 2022
Förkortad titelICPR 2022

Ämnesklassifikation (UKÄ)

  • Datorseende och robotik (autonoma system)


Utforska forskningsämnen för ”Points to patches: Enabling the use of self-attention for 3D shape recognition”. Tillsammans bildar de ett unikt fingeravtryck.

Citera det här