Image Classification Performance Evaluation for 3D Model Reconstruction

被引:0
|
作者
Yuniarti, Anny [1 ]
Suciati, Nanik [1 ]
Arifin, Agus Zainal [1 ]
机构
[1] Inst Teknol Sepuluh Nopember, Dept Informat, Surabaya, Indonesia
来源
2020 INTERNATIONAL CONFERENCE ON RADAR, ANTENNA, MICROWAVE, ELECTRONICS, AND TELECOMMUNICATIONS (ICRAMET): FOSTERING INNOVATION THROUGH ICTS FOR SUSTAINABLE SMART SOCIETY | 2020年
关键词
image classification; deep learning; 3D dataset; ShapeNet; rendering images;
D O I
10.1109/icramet51080.2020.9298643
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
3D reconstruction of 2D images is a classical problem in computer vision. Conventional methods have been proposed using multiple image registration, intrinsic and extrinsic camera parameter estimation, and optimization methods. Recently, the availability of a 3D dataset publicly shared has encouraged a deep-learning-based method for single-view reconstruction. One approach was by employing direct image encoding to the 3D point decoding approach as in PointNet and AtlasNet. Some other research directions attempted to retrieve 3D data using deep-learning-based methods, which employed a convolutional neural network (CNN). However, the use of CNN in image classification specific for the 3D reconstruction task still needs to be investigated because, usually, CNN was used as an image encoder in an auto-encoder setting instead of a classification module in a point generation network. Moreover, there is a lack of reports on the performance evaluation of deep-learning-based method on images rendered from 3D data, such as ShapeNet rendering images. In this paper, we implemented several deep-learning models to decode the ShapeNet rendering images that contain 13 model categories to examine the various hyper-parameters' impacts on each 3D model category. Our experiments showed that the hyper-parameters of the learning rate and epochs set to either 0.001 or 0.0001 and 60-80 epochs significantly outperformed other settings. Moreover, we observed that regardless of network configuration, some categories (plane, watercraft, car) performed better throughout the study. Therefore, a 3D reconstruction based on image classification can be designed based on the best performing categories.
引用
收藏
页码:229 / 233
页数:5
相关论文
共 50 条
  • [41] 3D EIT image reconstruction with GREIT
    Grychtol, Bartlomiej
    Mueller, Beat
    Adler, Andy
    PHYSIOLOGICAL MEASUREMENT, 2016, 37 (06) : 785 - 800
  • [42] Evaluation of MAP image reconstruction with positron range modeling for 3D PET
    Bai, Bing
    Laforest, Richard
    Smith, Anne M.
    Leahy, Richard M.
    2005 IEEE NUCLEAR SCIENCE SYMPOSIUM CONFERENCE RECORD, VOLS 1-5, 2005, : 2686 - 2689
  • [43] Quantitative NDT by 3D image reconstruction
    Schmitz, V
    Kroning, M
    Kangenberg, KJ
    ACOUSTICAL IMAGING, VOL 22, 1996, 22 : 735 - 744
  • [44] Image alignment for 3D reconstruction in a SEM
    Pintus, R
    Podda, S
    Vanzi, M
    MICROELECTRONICS RELIABILITY, 2005, 45 (9-11) : 1581 - 1584
  • [45] 3D Reconstruction from A Single Image
    Ping, Guiju
    Wang, Han
    PROCEEDINGS OF THE IEEE 2019 9TH INTERNATIONAL CONFERENCE ON CYBERNETICS AND INTELLIGENT SYSTEMS (CIS) ROBOTICS, AUTOMATION AND MECHATRONICS (RAM) (CIS & RAM 2019), 2019, : 47 - 52
  • [46] Error estimation of 3D reconstruction in 3D digital image correlation
    Zhu, Chengpeng
    Yu, Shanshan
    Liu, Cong
    Jiang, Pengfei
    Shao, Xinxing
    He, Xiaoyuan
    MEASUREMENT SCIENCE AND TECHNOLOGY, 2019, 30 (02)
  • [47] Passive 3D Face Reconstruction with 3D Digital Image Correlation
    Kieu, Hien
    Wang, Zhaoyang
    Le, Minh
    Nguyen, Hieu
    ADVANCEMENT OF OPTICAL METHODS IN EXPERIMENTAL MECHANICS, VOL 3, 2015, : 403 - 407
  • [48] Image segmentation and image matching for 3D terrain reconstruction
    Lu, YH
    Kubik, K
    Bennamoun, M
    FOURTEENTH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOLS 1 AND 2, 1998, : 1535 - 1537
  • [49] Model of vessels for 3D reconstruction
    León, LA
    Rodríguez, KB
    Hernández, CC
    Monreal, JE
    Pérez, JS
    METMBS'03: PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON MATHEMATICS AND ENGINEERING TECHNIQUES IN MEDICINE AND BIOLOGICAL SCIENCES, 2003, : 410 - 416
  • [50] Big Data Image Processing Based on Coefficient 3D Reconstruction Model
    Sun, Hujun
    MOBILE INFORMATION SYSTEMS, 2022, 2022