Human Shape Reconstruction with Loose Clothes from Partially Observed Data by Pose Specific Deformation

Akihiko Sayo, Hayato Onizuka, Diego Thomas, Yuta Nakashima, Hiroshi Kawasaki, Katsushi Ikeuchi

研究成果: Chapter in Book/Report/Conference proceedingConference contribution

抜粋

Reconstructing the entire body of moving human in a computer is important for various applications, such as tele-presence, virtual try-on, etc. For the purpose, realistic representation of loose clothes or non-rigid body deformation is a challenging and important task. Recent approaches for full-body reconstruction use a statistical shape model, which is built upon accurate full-body scans of people in skin-tight clothes. Such a model can be fitted to a point cloud of a person wearing loose clothes, however, it cannot represent the detailed shape of loose clothes, such as wrinkles and/or folds. In this paper, we propose a method that reconstructs 3D model of full-body human with loose clothes by reproducing the deformations as displacements from the skin-tight body mesh. We take advantage of a statistical shape model as base shape of full-body human mesh, and then, obtain displacements from the base mesh by non-rigid registration. To efficiently represent such displacements, we use lower dimensional embeddings of the deformations. This enables us to regress the coefficients corresponding to the small number of bases. We also propose a method to reconstruct shape only from a single 3D scanner, which is realized by shape fitting to only visible meshes as well as intra-frame shape interpolation. Our experiments with both unknown scene and partial body scans confirm the reconstruction ability of our proposed method.

元の言語英語
ホスト出版物のタイトルImage and Video Technology - 9th Pacific-Rim Symposium, PSIVT 2019, Proceedings
編集者Chilwoo Lee, Zhixun Su, Akihiro Sugimoto
出版者Springer
ページ225-239
ページ数15
ISBN(印刷物)9783030348786
DOI
出版物ステータス出版済み - 1 1 2019
イベント9th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2019 - Sydney, オーストラリア
継続期間: 11 18 201911 22 2019

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
11854 LNCS
ISSN(印刷物)0302-9743
ISSN(電子版)1611-3349

会議

会議9th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2019
オーストラリア
Sydney
期間11/18/1911/22/19

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

フィンガープリント Human Shape Reconstruction with Loose Clothes from Partially Observed Data by Pose Specific Deformation' の研究トピックを掘り下げます。これらはともに一意のフィンガープリントを構成します。

  • これを引用

    Sayo, A., Onizuka, H., Thomas, D., Nakashima, Y., Kawasaki, H., & Ikeuchi, K. (2019). Human Shape Reconstruction with Loose Clothes from Partially Observed Data by Pose Specific Deformation. : C. Lee, Z. Su, & A. Sugimoto (版), Image and Video Technology - 9th Pacific-Rim Symposium, PSIVT 2019, Proceedings (pp. 225-239). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 巻数 11854 LNCS). Springer. https://doi.org/10.1007/978-3-030-34879-3_18