Frankenstein3d: human body reconstruction from limited number of points

Download

index.pdf

Date

2019

Author

Taştan, Oğuzhan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

97
views

49
downloads

We propose a novel approach for reconstructing high-resolution 3D human body models from extremely small number of 3D points which represent body parts. We leverage a database of high-resolution 3D models of 100 humans varying from each other by physical attributes such as age, weight, size etc. We, first, divide the bodies in database into seven semantic regions. Then, for each input region consisting of maximum 40 points, we search the database for the best matching body part. For the matching criterion, we use the distance between novel point-base features of input points and body parts in the database. We further combine the matched parts from different bodies into one body which result in a high-resolution human body, with the help of Laplacian deformation. To evaluate our results, we pick points from each part of the ground-truth human body models, then reconstruct them using our method and compare the resulting bodies with corresponding ground-truths. Also, our results are compared with ARAP-based results. In addition, we run our algorithm with noisy data to test robustness of our method and run it with input points whose body parts are manually edited, which produces plausible human bodies that do not even exist in our database. Our experiments verify qualitatively and quantitatively that the proposed approach reconstructs human bodies with different physical attributes from small number of points successfully and prove that our method is robust to noisy data.

Subject Keywords

Modelling., Keywords: 3D Body Reconstruction, Modelling, Database-Assisted Modelling.

URI

http://etd.lib.metu.edu.tr/upload/12625045/index.pdf
https://hdl.handle.net/11511/45354

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Human body reconstruction from limited number of points Tastan, Oguzhan; Sahillioğlu, Yusuf (2021-04-01) We propose a novel approach for reconstructing plausible three-dimensional (3D) human body models from small number of 3D points which represent body parts. We leverage a database of 3D models of humans varying from each other by physical attributes such as age, gender, weight, and height. First we divide the bodies in database into seven semantic regions. Then, for each input region consisting of maximum 40 points, we search the database for the best matching body part. For the matching criterion, we use t...
Self-Supervised Learning of 3D Human Pose using Multi-view Geometry Kocabas, Muhammed; Karagoz, Salih; Akbaş, Emre (2019-01-01) Training accurate 3D human pose estimators requires large amount of 3D ground-truth data which is costly to collect. Various weakly or self supervised pose estimation methods have been proposed due to lack of 3D data. Nevertheless, these methods, in addition to 2D ground-truth poses, require either additional supervision in various forms (e.g. unpaired 3D ground truth data, a small subset of labels) or the camera parameters in multiview settings. To address these problems, we present EpipolarPose, a self-su...
Architectural Design in Virtual Reality and Mixed Reality Environments: A Comparative Analysis Ergün, Oğuzcan; Akın, Şahin; Gürsel Dino, İpek; Sürer, Elif (2019-01-01) Virtual reality (VR) provides a completely digital world of interaction which enables the users to modify, edit, and transform digital elements in a responsive way. Mixed reality (MR), which is the result of blending the digital world and the physical world together, brings new advancements and challenges to human, computer and environment interactions. This paper focuses on adapting the already-existing methods and tools in architecture to both VR and MR environments under sustainable architectural design ...
Nanobiomaterials: a review of the existing science and technology, and new approaches Hasırcı, Vasıf Nejat; Zorlutuna, P.; Ndreu, A.; Yilgor, P.; Basmanav, F. B.; Aydin, E. (2006-01-01) Nanotechnology has made great strides forward in the creation of new surfaces, new materials and new forms which also find application in the biomedical field. Traditional biomedical applications started benefiting from the use nanotechnology in an array of areas, such as biosensors, tissue engineering, controlled release systems, intelligent systems and nanocomposites used in implant design. In this manuscript a review of developments in these areas will be provided along with some applications from our la...
Comparison of 3D Versus 4D Path Planning for Unmanned Aerial Vehicles Cicibas, Halil; Demir, Kadir Alpaslan; ARICA, NAFİZ (2016-11-01) This research compares 3D versus 4D (three spatial dimensions and the time dimension) multi-objective and multi-criteria path-planning for unmanned aerial vehicles in complex dynamic environments. In this study, we empirically analyse the performances of 3D and 4D path planning approaches. Using the empirical data, we show that the 4D approach is superior over the 3D approach especially in complex dynamic environments. The research model consisting of flight objectives and criteria is developed based on int...

Citation Formats

O. Taştan, “Frankenstein3d: human body reconstruction from limited number of points,” Thesis (M.S.) -- Graduate School of Natural and Applied Sciences. Computer Engineering., Middle East Technical University, 2019.