SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation

Dingding Cai; Janne Heikkila; Esa Rahtu

doi:10.1109/3DV57658.2022.00065

SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation

Dingding Cai
, Janne Heikkila
, Esa Rahtu

Tietotekniikka

Tutkimustuotos: Konferenssiartikkeli › Tieteellinen › vertaisarvioitu

16 Sitaatiot (Scopus)

42 Lataukset (Pure)

Abstrakti

This paper presents an efficient symmetry-agnostic and correspondence-free framework, referred to as SC6D, for 6D object pose estimation from a single monocular RGB image. SC6D requires neither the 3D CAD model of the object nor any prior knowledge of the symmetries. The pose estimation is decomposed into three sub-tasks: a) object 3D rotation representation learning and matching; b) estimation of the 2D location of the object center; and c) scale-invariant distance estimation (the translation along the z-axis) via classification. SC6D is evaluated on three benchmark datasets, T-LESS, YCB-V, and ITODD, and results in state-of-the-art performance on the T-LESS dataset. More-over, SC6D is computationally much more efficient than the previous state-of-the-art method SurfEmb. The implementation and pre-trained models are publicly available at https://github.com/dingdingcai/SC6D-pose.

Alkuperäiskieli	Englanti
Otsikko	Proceedings - 2022 International Conference on 3D Vision, 3DV 2022
Kustantaja	IEEE
Sivut	536-546
Sivumäärä	11
ISBN (elektroninen)	978-1-6654-5670-8
DOI - pysyväislinkit	https://doi.org/10.1109/3DV57658.2022.00065
Tila	Julkaistu - 2022
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	International Conference on 3D Vision - Prague, Tshekki Kesto: 12 syysk. 2022 → 15 syysk. 2022

Julkaisusarja

Nimi	Proceedings - 2022 International Conference on 3D Vision, 3DV 2022
ISSN (elektroninen)	2475-7888

Conference

Conference	International Conference on 3D Vision
Maa/Alue	Tshekki
Kaupunki	Prague
Ajanjakso	12/09/22 → 15/09/22

Rahoitus

This work was supported by the Academy of Finland under the project #327910.

Julkaisufoorumi-taso

Jufo-taso 1

!!ASJC Scopus subject areas

Artificial Intelligence
Computer Vision and Pattern Recognition
Signal Processing

Pääsy asiakirjaan

10.1109/3DV57658.2022.00065

Symmetry-agnostic and Correspondence-free
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Accepted author manuscript, 9,03 MBLisenssi: Unspecified

https://urn.fi/URN:NBN:fi:tuni-202305024922Lisenssi: Unspecified

Siteeraa tätä

@inproceedings{de8d03e87b464128bb587f5d90acc3e0,

title = "SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation",

abstract = "This paper presents an efficient symmetry-agnostic and correspondence-free framework, referred to as SC6D, for 6D object pose estimation from a single monocular RGB image. SC6D requires neither the 3D CAD model of the object nor any prior knowledge of the symmetries. The pose estimation is decomposed into three sub-tasks: a) object 3D rotation representation learning and matching; b) estimation of the 2D location of the object center; and c) scale-invariant distance estimation (the translation along the z-axis) via classification. SC6D is evaluated on three benchmark datasets, T-LESS, YCB-V, and ITODD, and results in state-of-the-art performance on the T-LESS dataset. More-over, SC6D is computationally much more efficient than the previous state-of-the-art method SurfEmb. The implementation and pre-trained models are publicly available at https://github.com/dingdingcai/SC6D-pose.",

keywords = "6D object pose, correspondence free, symmetry agnostic",

author = "Dingding Cai and Janne Heikkila and Esa Rahtu",

note = "Funding Information: This work was supported by the Academy of Finland under the project \#327910. Publisher Copyright: {\textcopyright} 2022 IEEE.; International Conference on 3D Vision ; Conference date: 12-09-2022 Through 15-09-2022",

year = "2022",

doi = "10.1109/3DV57658.2022.00065",

language = "English",

series = "Proceedings - 2022 International Conference on 3D Vision, 3DV 2022",

publisher = "IEEE",

pages = "536--546",

booktitle = "Proceedings - 2022 International Conference on 3D Vision, 3DV 2022",

address = "United States",

}

Cai, D, Heikkila, J & Rahtu, E 2022, SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation. julkaisussa Proceedings - 2022 International Conference on 3D Vision, 3DV 2022. Proceedings - 2022 International Conference on 3D Vision, 3DV 2022, IEEE, Sivut 536-546, International Conference on 3D Vision, Prague, Tshekki, 12/09/22. https://doi.org/10.1109/3DV57658.2022.00065

TY - GEN

T1 - SC6D

T2 - International Conference on 3D Vision

AU - Cai, Dingding

AU - Heikkila, Janne

AU - Rahtu, Esa

PY - 2022

Y1 - 2022

N2 - This paper presents an efficient symmetry-agnostic and correspondence-free framework, referred to as SC6D, for 6D object pose estimation from a single monocular RGB image. SC6D requires neither the 3D CAD model of the object nor any prior knowledge of the symmetries. The pose estimation is decomposed into three sub-tasks: a) object 3D rotation representation learning and matching; b) estimation of the 2D location of the object center; and c) scale-invariant distance estimation (the translation along the z-axis) via classification. SC6D is evaluated on three benchmark datasets, T-LESS, YCB-V, and ITODD, and results in state-of-the-art performance on the T-LESS dataset. More-over, SC6D is computationally much more efficient than the previous state-of-the-art method SurfEmb. The implementation and pre-trained models are publicly available at https://github.com/dingdingcai/SC6D-pose.

AB - This paper presents an efficient symmetry-agnostic and correspondence-free framework, referred to as SC6D, for 6D object pose estimation from a single monocular RGB image. SC6D requires neither the 3D CAD model of the object nor any prior knowledge of the symmetries. The pose estimation is decomposed into three sub-tasks: a) object 3D rotation representation learning and matching; b) estimation of the 2D location of the object center; and c) scale-invariant distance estimation (the translation along the z-axis) via classification. SC6D is evaluated on three benchmark datasets, T-LESS, YCB-V, and ITODD, and results in state-of-the-art performance on the T-LESS dataset. More-over, SC6D is computationally much more efficient than the previous state-of-the-art method SurfEmb. The implementation and pre-trained models are publicly available at https://github.com/dingdingcai/SC6D-pose.

KW - 6D object pose

KW - correspondence free

KW - symmetry agnostic

U2 - 10.1109/3DV57658.2022.00065

DO - 10.1109/3DV57658.2022.00065

M3 - Conference contribution

AN - SCOPUS:85149285578

T3 - Proceedings - 2022 International Conference on 3D Vision, 3DV 2022

SP - 536

EP - 546

BT - Proceedings - 2022 International Conference on 3D Vision, 3DV 2022

PB - IEEE

Y2 - 12 September 2022 through 15 September 2022

ER -

SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation

Abstrakti

Julkaisusarja

Conference

Rahoitus

Julkaisufoorumi-taso

!!ASJC Scopus subject areas

Pääsy asiakirjaan

Sormenjälki

Siteeraa tätä