Cross-Domain Indoor Visual Place Recognition for Mobile Robot via Generalization Using Style Augmentation

被引:3
作者
Wozniak, Piotr [1 ]
Ozog, Dominik [1 ]
机构
[1] Rzeszow Univ Technol, Fac Elect & Comp Engn, Dept Comp & Control Engn, Al Powstancow Warszawy 12, PL-35959 Rzeszow, Poland
关键词
visual place recognition; CNNs; multi-domain learning; domain generalization; transfer learning;
D O I
10.3390/s23136134
中图分类号
O65 [分析化学];
学科分类号
070302 ; 081704 ;
摘要
The article presents an algorithm for the multi-domain visual recognition of an indoor place. It is based on a convolutional neural network and style randomization. The authors proposed a scene classification mechanism and improved the performance of the models based on synthetic and real data from various domains. In the proposed dataset, a domain change was defined as a camera model change. A dataset of images collected from several rooms was used to show different scenarios, human actions, equipment changes, and lighting conditions. The proposed method was tested in a scene classification problem where multi-domain data were used. The basis was a transfer learning approach with an extension style applied to various combinations of source and target data. The focus was on improving the unknown domain score and multi-domain support. The results of the experiments were analyzed in the context of data collected on a humanoid robot. The article shows that the average score was the highest for the use of multi-domain data and data style enhancement. The method of obtaining average results for the proposed method reached the level of 92.08%. The result obtained by another research team was corrected.
引用
收藏
页数:19
相关论文
共 53 条
  • [11] Deng J, 2009, PROC CVPR IEEE, P248, DOI 10.1109/CVPRW.2009.5206848
  • [12] A CNN-Based System for Mobile Robot Navigation in Indoor Environments via Visual Localization with a Small Dataset
    Foroughi, Farzin
    Chen, Zonghai
    Wang, Jikai
    [J]. WORLD ELECTRIC VEHICLE JOURNAL, 2021, 12 (03):
  • [13] Visual simultaneous localization and mapping: a survey
    Fuentes-Pacheco, Jorge
    Ruiz-Ascencio, Jose
    Manuel Rendon-Mancha, Juan
    [J]. ARTIFICIAL INTELLIGENCE REVIEW, 2015, 43 (01) : 55 - 81
  • [14] An Efficient Object Navigation Strategy for Mobile Robots Based on Semantic Information
    Guo, Yu
    Xie, Yuanyan
    Chen, Yue
    Ban, Xiaojuan
    Sadoun, Balqies
    Obaidat, Mohammad S.
    [J]. ELECTRONICS, 2022, 11 (07)
  • [15] Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition
    Hausler, Stephen
    Garg, Sourav
    Xu, Ming
    Milford, Michael
    Fischer, Tobias
    [J]. 2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 14136 - 14147
  • [16] He KM, 2015, Arxiv, DOI arXiv:1512.03385
  • [17] On evaluation metrics for medical applications of artificial intelligence
    Hicks, Steven A.
    Struemke, Inga
    Thambawita, Vajira
    Hammou, Malek
    Riegler, Michael A.
    Halvorsen, Pal
    Parasa, Sravanthi
    [J]. SCIENTIFIC REPORTS, 2022, 12 (01)
  • [18] Transfer learning: a friendly introduction
    Hosna, Asmaul
    Merry, Ethel
    Gyalmo, Jigmey
    Alom, Zulfikar
    Aung, Zeyar
    Azim, Mohammad Abdul
    [J]. JOURNAL OF BIG DATA, 2022, 9 (01)
  • [19] Become Competent within One Day in Generating Boxplots and Violin Plots for a Novice without Prior R Experience
    Hu, Kejin
    [J]. METHODS AND PROTOCOLS, 2020, 3 (04) : 1 - 30
  • [20] Inoue N, 2018, Arxiv, DOI arXiv:1803.11365