Shuffle in 3D: A Lightweight Architecture for Stereo Matching

The deep learning-based stereo matching approaches commonly construct 3D cost volume with a Siamese network, and the 3D encoder-decoder architectures regularize 3D cost volume to generate disparity as output. Yet, the 3D encoder-decoder architectures produce tremendous parameters that causing huge t...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) s. 675 - 677
Hlavní autoři: Xiao, Jianqiang, Yamane, Satoshi
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 12.10.2021
Témata:
3D ShuffleNet
Computational efficiency
Computer architecture
Conferences
Consumer electronics
Costs
lightweight 3D encoder-decoder
Stereo Matching
Three-dimensional displays
Training
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The deep learning-based stereo matching approaches commonly construct 3D cost volume with a Siamese network, and the 3D encoder-decoder architectures regularize 3D cost volume to generate disparity as output. Yet, the 3D encoder-decoder architectures produce tremendous parameters that causing huge training and inference cost. In this paper, we propose a 3D ShuffleNet-based approach on PSMNet, which greatly saves the computational cost without losing without reducing the performance of the network.
DOI:10.1109/GCCE53005.2021.9622083