Part-aware Panoptic Segmentation

Daan de Geus; Panagiotis Meletis; Chenyang Lu; Xiaoxiao Wen; Gijs Dubbelman

doi:10.1109/CVPR46437.2021.00544

Part-aware Panoptic Segmentation

Daan de Geus, Panagiotis Meletis, Chenyang Lu, Xiaoxiao Wen, Gijs Dubbelman

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

42 Citations (Scopus)

93 Downloads (Pure)

Abstract

In this work, we introduce the new scene understanding task of Part-aware Panoptic Segmentation (PPS), which aims to understand a scene at multiple levels of abstraction, and unifies the tasks of scene parsing and part parsing. For this novel task, we provide consistent annotations on two commonly used datasets: Cityscapes and Pascal VOC. Moreover, we present a single metric to evaluate PPS, called Part-aware Panoptic Quality (PartPQ). For this new task, using the metric and annotations, we set multiple baselines by merging results of existing state-of-the-art methods for panoptic segmentation and part segmentation. Finally, we conduct several experiments that evaluate the importance of the different levels of abstraction in this single task.

Original language	English
Title of host publication	2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher	Institute of Electrical and Electronics Engineers
Pages	5481-5490
Number of pages	10
ISBN (Electronic)	978-1-6654-4509-2
DOIs	https://doi.org/10.1109/CVPR46437.2021.00544
Publication status	Published - 13 Nov 2021
Event	2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition - Duration: 19 Jun 2021 → 25 Jun 2021

Conference

Conference	2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition
Abbreviated title	CVPR
Period	19/06/21 → 25/06/21

Access to Document

10.1109/CVPR46437.2021.00544

published versionFinal published version, 3.23 MB

https://openaccess.thecvf.com/content/CVPR2021/html/de_Geus_Part-Aware_Panoptic_Segmentation_CVPR_2021_paper.html

Cite this

@inproceedings{5ce5021d47754956acd89df20d6b461d,

title = "Part-aware Panoptic Segmentation",

abstract = "In this work, we introduce the new scene understanding task of Part-aware Panoptic Segmentation (PPS), which aims to understand a scene at multiple levels of abstraction, and unifies the tasks of scene parsing and part parsing. For this novel task, we provide consistent annotations on two commonly used datasets: Cityscapes and Pascal VOC. Moreover, we present a single metric to evaluate PPS, called Part-aware Panoptic Quality (PartPQ). For this new task, using the metric and annotations, we set multiple baselines by merging results of existing state-of-the-art methods for panoptic segmentation and part segmentation. Finally, we conduct several experiments that evaluate the importance of the different levels of abstraction in this single task.",

author = "{de Geus}, Daan and Panagiotis Meletis and Chenyang Lu and Xiaoxiao Wen and Gijs Dubbelman",

year = "2021",

month = nov,

day = "13",

doi = "10.1109/CVPR46437.2021.00544",

language = "English",

pages = "5481--5490",

booktitle = "2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR ; Conference date: 19-06-2021 Through 25-06-2021",

}

de Geus, D, Meletis, P, Lu, C, Wen, X & Dubbelman, G 2021, Part-aware Panoptic Segmentation. in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)., 9578144, Institute of Electrical and Electronics Engineers, pp. 5481-5490, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 19/06/21. https://doi.org/10.1109/CVPR46437.2021.00544

TY - GEN

T1 - Part-aware Panoptic Segmentation

AU - de Geus, Daan

AU - Meletis, Panagiotis

AU - Lu, Chenyang

AU - Wen, Xiaoxiao

AU - Dubbelman, Gijs

PY - 2021/11/13

Y1 - 2021/11/13

N2 - In this work, we introduce the new scene understanding task of Part-aware Panoptic Segmentation (PPS), which aims to understand a scene at multiple levels of abstraction, and unifies the tasks of scene parsing and part parsing. For this novel task, we provide consistent annotations on two commonly used datasets: Cityscapes and Pascal VOC. Moreover, we present a single metric to evaluate PPS, called Part-aware Panoptic Quality (PartPQ). For this new task, using the metric and annotations, we set multiple baselines by merging results of existing state-of-the-art methods for panoptic segmentation and part segmentation. Finally, we conduct several experiments that evaluate the importance of the different levels of abstraction in this single task.

AB - In this work, we introduce the new scene understanding task of Part-aware Panoptic Segmentation (PPS), which aims to understand a scene at multiple levels of abstraction, and unifies the tasks of scene parsing and part parsing. For this novel task, we provide consistent annotations on two commonly used datasets: Cityscapes and Pascal VOC. Moreover, we present a single metric to evaluate PPS, called Part-aware Panoptic Quality (PartPQ). For this new task, using the metric and annotations, we set multiple baselines by merging results of existing state-of-the-art methods for panoptic segmentation and part segmentation. Finally, we conduct several experiments that evaluate the importance of the different levels of abstraction in this single task.

UR - http://www.scopus.com/inward/record.url?scp=85117894544&partnerID=8YFLogxK

U2 - 10.1109/CVPR46437.2021.00544

DO - 10.1109/CVPR46437.2021.00544

M3 - Conference contribution

SP - 5481

EP - 5490

BT - 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

PB - Institute of Electrical and Electronics Engineers

T2 - 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition

Y2 - 19 June 2021 through 25 June 2021

ER -

Part-aware Panoptic Segmentation

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this