Twins-Mix: Self Mixing in Latent Space for Reasonable Data Augmentation of 3D Computer-Aided Design Generative Modeling

被引:1
作者
Li, Xueyang [1 ]
Xu, Minyang [1 ]
Zhou, Xiangdong [1 ]
机构
[1] Fudan Univ, Sch Comp Sci, Shanghai, Peoples R China
来源
2023 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, ICME | 2023年
关键词
CAD generative modeling; data augmentation; mixup;
D O I
10.1109/ICME55011.2023.00160
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep generative modeling on parametric Computer-Aided Design (CAD) data has great potentials in various industrial scenarios. However, the reasonable data augmentation for parametric CAD has not been solved yet, which limits the performance of the deep learning on parametric CAD data. Unlike images or language, parametric CAD data involves dual-aspects: command sequence and geometric shape. Hence, most previous data augmentation methods are unsuitable for this problem. To address this issue, we propose a novel mix-based augmentation method, namely Twins-Mix, keeping a well balance between the diversity and the validity of parametric CAD sequences, which significantly boost the performance of CAD generative modeling. Comprehensive experiments are conducted on the commonly used benchmark datasets, i.e., Fusion 360 and DeepCAD. The experimental results demonstrate that our model exceeds other comparable augmentations on CAD generative modeling significantly, especially in increasing the valid 3D shape construction ratio by at least more than 4%.
引用
收藏
页码:906 / 911
页数:6
相关论文
共 27 条
  • [1] Arjovsky M, 2017, PR MACH LEARN RES, V70
  • [2] Bengio Y., 2013, PMLR, P552
  • [3] ReMix: Towards Image-to-Image Translation with Limited Data
    Cao, Jie
    Hou, Luanxuan
    Yang, Ming-Hsuan
    He, Ran
    Sun, Zhenan
    [J]. 2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 15013 - 15022
  • [4] Cao Weijuan, 2020, INT DES ENG TECHN C, V84003
  • [5] Carlier A., 2020, ADV NEURAL INFORM PR, V33, P16351, DOI DOI 10.5555/3495724.3497096
  • [6] GeoAug: Data Augmentation for Few-Shot NeRF with Geometry Constraints
    Chen, Di
    Liu, Yu
    Huang, Lianghua
    Wang, Bin
    Pan, Pan
    [J]. COMPUTER VISION - ECCV 2022, PT XVII, 2022, 13677 : 322 - 337
  • [7] Chen JA, 2020, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020), P2147
  • [8] SuperMix: Supervising the Mixing Data Augmentation
    Dabouei, Ali
    Soleymani, Sobhan
    Taherkhani, Fariborz
    Nasrabadi, Nasser M.
    [J]. 2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 13789 - 13798
  • [9] Ganin Y., 2021, Adv. Neural Inf. Proces. Syst., V34, P5885
  • [10] Deep Residual Learning for Image Recognition
    He, Kaiming
    Zhang, Xiangyu
    Ren, Shaoqing
    Sun, Jian
    [J]. 2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, : 770 - 778