Extended 2D Scene Sketch-Based 3D Scene Retrieval

J. Yuan, H. Abdul-Rashid, B. Li, Y. Lu, T. Schreck, N.-M. Bui, T.-L. Do, K.-T. Nguyen, T.-A. Nguyen, V.-T.- Nguyen, M.-T. Tran, T. Wang

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in einem KonferenzbandForschungBegutachtung

Abstract

Sketch-based 3D scene retrieval is to retrieve 3D scene models given a user's hand-drawn 2D scene sketch. It is a brand new but also very challenging research topic in the field of 3D object retrieval due to the semantic gap in their representations: 3D scene models or views differ from non-realistic 2D scene sketches. To boost this interesting research, we organized a 2D Scene Sketch-Based 3D Scene Retrieval track in SHREC'18, resulting a SceneSBR18 benchmark which contains 10 scene classes. In order to make it more comprehensive, we have extended the number of the scene categories from the initial 10 classes in the SceneSBR2018 benchmark to 30 classes, resulting in a new and more challenging benchmark SceneSBR2019 which has 750 2D scene sketches and 3,000 3D scene models. Therefore, the objective of this track is to further evaluate the performance and scalability of different 2D scene sketch-based 3D scene model retrieval algorithms using this extended and more comprehensive new benchmark. In this track, two groups from USA and Vietnam have successfully submitted 4 runs. Based on 7 commonly used retrieval metrics, we evaluate their retrieval performance. We have also conducted a comprehensive analysis and discussion of these methods and proposed several future research directions to deal with this challenging research topic. Deep learning techniques have been proved their great potentials again in dealing with this challenging retrieval task, in terms of both retrieval accuracy and scalability to a larger dataset. We hope this publicly available benchmark, together with its evaluation results and source code, will further enrich and promote 2D scene sketch-based 3D scene retrieval research area and its corresponding applications.
Originalspracheenglisch
TitelEurographics Workshop on 3D Object Retrieval
Redakteure/-innenSilvia Biasotti, Guillaume Lavoue, Remco Veltkamp
Herausgeber (Verlag)Eurographics - European Association for Computer Graphics
Seiten33-39
Seitenumfang7
ISBN (Print)978-3-03868-077-2
DOIs
PublikationsstatusVeröffentlicht - 2019
Veranstaltung12th Eurographics Workshop on 3D Object Retrieval - Genova, Italien
Dauer: 5 Mai 20196 Mai 2019
http://3dor2019.ge.imati.cnr.it/

Workshop

Workshop12th Eurographics Workshop on 3D Object Retrieval
LandItalien
OrtGenova
Zeitraum5/05/196/05/19
Internetadresse

Fingerprint

Scalability
Semantics
Deep learning

Fields of Expertise

  • Information, Communication & Computing

Dies zitieren

Yuan, J., Abdul-Rashid, H., Li, B., Lu, Y., Schreck, T., Bui, N-M., ... Wang, T. (2019). Extended 2D Scene Sketch-Based 3D Scene Retrieval. in S. Biasotti, G. Lavoue, & R. Veltkamp (Hrsg.), Eurographics Workshop on 3D Object Retrieval (S. 33-39). Eurographics - European Association for Computer Graphics. https://doi.org/10.2312/3dor.20191059

Extended 2D Scene Sketch-Based 3D Scene Retrieval. / Yuan, J.; Abdul-Rashid, H.; Li, B.; Lu, Y.; Schreck, T.; Bui, N.-M.; Do, T.-L.; Nguyen, K.-T.; Nguyen, T.-A.; Nguyen, V.-T.-; Tran, M.-T.; Wang, T.

Eurographics Workshop on 3D Object Retrieval. Hrsg. / Silvia Biasotti; Guillaume Lavoue; Remco Veltkamp. Eurographics - European Association for Computer Graphics, 2019. S. 33-39.

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in einem KonferenzbandForschungBegutachtung

Yuan, J, Abdul-Rashid, H, Li, B, Lu, Y, Schreck, T, Bui, N-M, Do, T-L, Nguyen, K-T, Nguyen, T-A, Nguyen, V-T, Tran, M-T & Wang, T 2019, Extended 2D Scene Sketch-Based 3D Scene Retrieval. in S Biasotti, G Lavoue & R Veltkamp (Hrsg.), Eurographics Workshop on 3D Object Retrieval. Eurographics - European Association for Computer Graphics, S. 33-39, Genova, Italien, 5/05/19. https://doi.org/10.2312/3dor.20191059
Yuan J, Abdul-Rashid H, Li B, Lu Y, Schreck T, Bui N-M et al. Extended 2D Scene Sketch-Based 3D Scene Retrieval. in Biasotti S, Lavoue G, Veltkamp R, Hrsg., Eurographics Workshop on 3D Object Retrieval. Eurographics - European Association for Computer Graphics. 2019. S. 33-39 https://doi.org/10.2312/3dor.20191059
Yuan, J. ; Abdul-Rashid, H. ; Li, B. ; Lu, Y. ; Schreck, T. ; Bui, N.-M. ; Do, T.-L. ; Nguyen, K.-T. ; Nguyen, T.-A. ; Nguyen, V.-T.- ; Tran, M.-T. ; Wang, T. / Extended 2D Scene Sketch-Based 3D Scene Retrieval. Eurographics Workshop on 3D Object Retrieval. Hrsg. / Silvia Biasotti ; Guillaume Lavoue ; Remco Veltkamp. Eurographics - European Association for Computer Graphics, 2019. S. 33-39
@inproceedings{5f0823949a05422c90c26c29d952f2e4,
title = "Extended 2D Scene Sketch-Based 3D Scene Retrieval",
abstract = "Sketch-based 3D scene retrieval is to retrieve 3D scene models given a user's hand-drawn 2D scene sketch. It is a brand new but also very challenging research topic in the field of 3D object retrieval due to the semantic gap in their representations: 3D scene models or views differ from non-realistic 2D scene sketches. To boost this interesting research, we organized a 2D Scene Sketch-Based 3D Scene Retrieval track in SHREC'18, resulting a SceneSBR18 benchmark which contains 10 scene classes. In order to make it more comprehensive, we have extended the number of the scene categories from the initial 10 classes in the SceneSBR2018 benchmark to 30 classes, resulting in a new and more challenging benchmark SceneSBR2019 which has 750 2D scene sketches and 3,000 3D scene models. Therefore, the objective of this track is to further evaluate the performance and scalability of different 2D scene sketch-based 3D scene model retrieval algorithms using this extended and more comprehensive new benchmark. In this track, two groups from USA and Vietnam have successfully submitted 4 runs. Based on 7 commonly used retrieval metrics, we evaluate their retrieval performance. We have also conducted a comprehensive analysis and discussion of these methods and proposed several future research directions to deal with this challenging research topic. Deep learning techniques have been proved their great potentials again in dealing with this challenging retrieval task, in terms of both retrieval accuracy and scalability to a larger dataset. We hope this publicly available benchmark, together with its evaluation results and source code, will further enrich and promote 2D scene sketch-based 3D scene retrieval research area and its corresponding applications.",
author = "J. Yuan and H. Abdul-Rashid and B. Li and Y. Lu and T. Schreck and N.-M. Bui and T.-L. Do and K.-T. Nguyen and T.-A. Nguyen and V.-T.- Nguyen and M.-T. Tran and T. Wang",
year = "2019",
doi = "10.2312/3dor.20191059",
language = "English",
isbn = "978-3-03868-077-2",
pages = "33--39",
editor = "Silvia Biasotti and Guillaume Lavoue and Remco Veltkamp",
booktitle = "Eurographics Workshop on 3D Object Retrieval",
publisher = "Eurographics - European Association for Computer Graphics",
address = "Switzerland",

}

TY - GEN

T1 - Extended 2D Scene Sketch-Based 3D Scene Retrieval

AU - Yuan, J.

AU - Abdul-Rashid, H.

AU - Li, B.

AU - Lu, Y.

AU - Schreck, T.

AU - Bui, N.-M.

AU - Do, T.-L.

AU - Nguyen, K.-T.

AU - Nguyen, T.-A.

AU - Nguyen, V.-T.-

AU - Tran, M.-T.

AU - Wang, T.

PY - 2019

Y1 - 2019

N2 - Sketch-based 3D scene retrieval is to retrieve 3D scene models given a user's hand-drawn 2D scene sketch. It is a brand new but also very challenging research topic in the field of 3D object retrieval due to the semantic gap in their representations: 3D scene models or views differ from non-realistic 2D scene sketches. To boost this interesting research, we organized a 2D Scene Sketch-Based 3D Scene Retrieval track in SHREC'18, resulting a SceneSBR18 benchmark which contains 10 scene classes. In order to make it more comprehensive, we have extended the number of the scene categories from the initial 10 classes in the SceneSBR2018 benchmark to 30 classes, resulting in a new and more challenging benchmark SceneSBR2019 which has 750 2D scene sketches and 3,000 3D scene models. Therefore, the objective of this track is to further evaluate the performance and scalability of different 2D scene sketch-based 3D scene model retrieval algorithms using this extended and more comprehensive new benchmark. In this track, two groups from USA and Vietnam have successfully submitted 4 runs. Based on 7 commonly used retrieval metrics, we evaluate their retrieval performance. We have also conducted a comprehensive analysis and discussion of these methods and proposed several future research directions to deal with this challenging research topic. Deep learning techniques have been proved their great potentials again in dealing with this challenging retrieval task, in terms of both retrieval accuracy and scalability to a larger dataset. We hope this publicly available benchmark, together with its evaluation results and source code, will further enrich and promote 2D scene sketch-based 3D scene retrieval research area and its corresponding applications.

AB - Sketch-based 3D scene retrieval is to retrieve 3D scene models given a user's hand-drawn 2D scene sketch. It is a brand new but also very challenging research topic in the field of 3D object retrieval due to the semantic gap in their representations: 3D scene models or views differ from non-realistic 2D scene sketches. To boost this interesting research, we organized a 2D Scene Sketch-Based 3D Scene Retrieval track in SHREC'18, resulting a SceneSBR18 benchmark which contains 10 scene classes. In order to make it more comprehensive, we have extended the number of the scene categories from the initial 10 classes in the SceneSBR2018 benchmark to 30 classes, resulting in a new and more challenging benchmark SceneSBR2019 which has 750 2D scene sketches and 3,000 3D scene models. Therefore, the objective of this track is to further evaluate the performance and scalability of different 2D scene sketch-based 3D scene model retrieval algorithms using this extended and more comprehensive new benchmark. In this track, two groups from USA and Vietnam have successfully submitted 4 runs. Based on 7 commonly used retrieval metrics, we evaluate their retrieval performance. We have also conducted a comprehensive analysis and discussion of these methods and proposed several future research directions to deal with this challenging research topic. Deep learning techniques have been proved their great potentials again in dealing with this challenging retrieval task, in terms of both retrieval accuracy and scalability to a larger dataset. We hope this publicly available benchmark, together with its evaluation results and source code, will further enrich and promote 2D scene sketch-based 3D scene retrieval research area and its corresponding applications.

U2 - 10.2312/3dor.20191059

DO - 10.2312/3dor.20191059

M3 - Conference contribution

SN - 978-3-03868-077-2

SP - 33

EP - 39

BT - Eurographics Workshop on 3D Object Retrieval

A2 - Biasotti, Silvia

A2 - Lavoue, Guillaume

A2 - Veltkamp, Remco

PB - Eurographics - European Association for Computer Graphics

ER -