HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction
Majority of current research focuses on a single static object reconstruction from a given pointcloud. However, the existing approaches are not applicable to real world applications such as dynamic and morphing scene reconstruction. To solve this, we propose a novel two-tiered deep neural network ar...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/12/3945 |
| _version_ | 1797531043525296128 |
|---|---|
| author | Audrius Kulikajevas Rytis Maskeliunas Robertas Damasevicius Rafal Scherer |
| author_facet | Audrius Kulikajevas Rytis Maskeliunas Robertas Damasevicius Rafal Scherer |
| author_sort | Audrius Kulikajevas |
| collection | DOAJ |
| description | Majority of current research focuses on a single static object reconstruction from a given pointcloud. However, the existing approaches are not applicable to real world applications such as dynamic and morphing scene reconstruction. To solve this, we propose a novel two-tiered deep neural network architecture, which is capable of reconstructing self-obstructed human-like morphing shapes from a depth frame in conjunction with cameras intrinsic parameters. The tests were performed using on custom dataset generated using a combination of AMASS and MoVi datasets. The proposed network achieved Jaccards’ Index of 0.7907 for the first tier, which is used to extract region of interest from the point cloud. The second tier of the network has achieved Earth Mover’s distance of 0.0256 and Chamfer distance of 0.276, indicating good experimental results. Further, subjective reconstruction results inspection shows strong predictive capabilities of the network, with the solution being able to reconstruct limb positions from very few object details. |
| first_indexed | 2024-03-10T10:37:34Z |
| format | Article |
| id | doaj.art-e2076f7edc7042fdba176d05b398547b |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T10:37:34Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-e2076f7edc7042fdba176d05b398547b2023-11-21T23:12:01ZengMDPI AGSensors1424-82202021-06-012112394510.3390/s21123945HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose ReconstructionAudrius Kulikajevas0Rytis Maskeliunas1Robertas Damasevicius2Rafal Scherer3Department of Multimedia Engineering, Kaunas University of Technology, 51368 Kaunas, LithuaniaDepartment of Multimedia Engineering, Kaunas University of Technology, 51368 Kaunas, LithuaniaFaculty of Applied Mathematics, Silesian University of Technology, 44-100 Gliwice, PolandDepartment of Intelligent Computer Systems, Częstochowa University of Technology, 42-200 Częstochowa, PolandMajority of current research focuses on a single static object reconstruction from a given pointcloud. However, the existing approaches are not applicable to real world applications such as dynamic and morphing scene reconstruction. To solve this, we propose a novel two-tiered deep neural network architecture, which is capable of reconstructing self-obstructed human-like morphing shapes from a depth frame in conjunction with cameras intrinsic parameters. The tests were performed using on custom dataset generated using a combination of AMASS and MoVi datasets. The proposed network achieved Jaccards’ Index of 0.7907 for the first tier, which is used to extract region of interest from the point cloud. The second tier of the network has achieved Earth Mover’s distance of 0.0256 and Chamfer distance of 0.276, indicating good experimental results. Further, subjective reconstruction results inspection shows strong predictive capabilities of the network, with the solution being able to reconstruct limb positions from very few object details.https://www.mdpi.com/1424-8220/21/12/39453D shape recognition3D depth scanningpointcloud reconstructionhuman shape reconstruction |
| spellingShingle | Audrius Kulikajevas Rytis Maskeliunas Robertas Damasevicius Rafal Scherer HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction Sensors 3D shape recognition 3D depth scanning pointcloud reconstruction human shape reconstruction |
| title | HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction |
| title_full | HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction |
| title_fullStr | HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction |
| title_full_unstemmed | HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction |
| title_short | HUMANNET—A Two-Tiered Deep Neural Network Architecture for Self-Occluding Humanoid Pose Reconstruction |
| title_sort | humannet a two tiered deep neural network architecture for self occluding humanoid pose reconstruction |
| topic | 3D shape recognition 3D depth scanning pointcloud reconstruction human shape reconstruction |
| url | https://www.mdpi.com/1424-8220/21/12/3945 |
| work_keys_str_mv | AT audriuskulikajevas humannetatwotiereddeepneuralnetworkarchitectureforselfoccludinghumanoidposereconstruction AT rytismaskeliunas humannetatwotiereddeepneuralnetworkarchitectureforselfoccludinghumanoidposereconstruction AT robertasdamasevicius humannetatwotiereddeepneuralnetworkarchitectureforselfoccludinghumanoidposereconstruction AT rafalscherer humannetatwotiereddeepneuralnetworkarchitectureforselfoccludinghumanoidposereconstruction |