Behind Every Domain There is a Shift: Adapting Distortion-aware Vision Transformers for Panoramic Semantic Segmentation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Jiaming, Yang, Kailun, Shi, Hao, Reiß, Simon, Peng, Kunyu, Ma, Chaoxiang, Fu, Haodong, Torr, Philip H. S., Wang, Kaiwei, Stiefelhagen, Rainer
Formato: Preprint
Publicado: 2022
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910464930217984
author Zhang, Jiaming
Yang, Kailun
Shi, Hao
Reiß, Simon
Peng, Kunyu
Ma, Chaoxiang
Fu, Haodong
Torr, Philip H. S.
Wang, Kaiwei
Stiefelhagen, Rainer
author_facet Zhang, Jiaming
Yang, Kailun
Shi, Hao
Reiß, Simon
Peng, Kunyu
Ma, Chaoxiang
Fu, Haodong
Torr, Philip H. S.
Wang, Kaiwei
Stiefelhagen, Rainer
contents In this paper, we address panoramic semantic segmentation which is under-explored due to two critical challenges: (1) image distortions and object deformations on panoramas; (2) lack of semantic annotations in the 360° imagery. To tackle these problems, first, we propose the upgraded Transformer for Panoramic Semantic Segmentation, i.e., Trans4PASS+, equipped with Deformable Patch Embedding (DPE) and Deformable MLP (DMLPv2) modules for handling object deformations and image distortions whenever (before or after adaptation) and wherever (shallow or deep levels). Second, we enhance the Mutual Prototypical Adaptation (MPA) strategy via pseudo-label rectification for unsupervised domain adaptive panoramic segmentation. Third, aside from Pinhole-to-Panoramic (Pin2Pan) adaptation, we create a new dataset (SynPASS) with 9,080 panoramic images, facilitating Synthetic-to-Real (Syn2Real) adaptation scheme in 360° imagery. Extensive experiments are conducted, which cover indoor and outdoor scenarios, and each of them is investigated with Pin2Pan and Syn2Real regimens. Trans4PASS+ achieves state-of-the-art performances on four domain adaptive panoramic semantic segmentation benchmarks. Code is available at https://github.com/jamycheung/Trans4PASS.
format Preprint
id arxiv_https___arxiv_org_abs_2207_11860
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Behind Every Domain There is a Shift: Adapting Distortion-aware Vision Transformers for Panoramic Semantic Segmentation
Zhang, Jiaming
Yang, Kailun
Shi, Hao
Reiß, Simon
Peng, Kunyu
Ma, Chaoxiang
Fu, Haodong
Torr, Philip H. S.
Wang, Kaiwei
Stiefelhagen, Rainer
Computer Vision and Pattern Recognition
Robotics
Image and Video Processing
In this paper, we address panoramic semantic segmentation which is under-explored due to two critical challenges: (1) image distortions and object deformations on panoramas; (2) lack of semantic annotations in the 360° imagery. To tackle these problems, first, we propose the upgraded Transformer for Panoramic Semantic Segmentation, i.e., Trans4PASS+, equipped with Deformable Patch Embedding (DPE) and Deformable MLP (DMLPv2) modules for handling object deformations and image distortions whenever (before or after adaptation) and wherever (shallow or deep levels). Second, we enhance the Mutual Prototypical Adaptation (MPA) strategy via pseudo-label rectification for unsupervised domain adaptive panoramic segmentation. Third, aside from Pinhole-to-Panoramic (Pin2Pan) adaptation, we create a new dataset (SynPASS) with 9,080 panoramic images, facilitating Synthetic-to-Real (Syn2Real) adaptation scheme in 360° imagery. Extensive experiments are conducted, which cover indoor and outdoor scenarios, and each of them is investigated with Pin2Pan and Syn2Real regimens. Trans4PASS+ achieves state-of-the-art performances on four domain adaptive panoramic semantic segmentation benchmarks. Code is available at https://github.com/jamycheung/Trans4PASS.
title Behind Every Domain There is a Shift: Adapting Distortion-aware Vision Transformers for Panoramic Semantic Segmentation
topic Computer Vision and Pattern Recognition
Robotics
Image and Video Processing
url https://arxiv.org/abs/2207.11860