MisConv: Convolutional Neural Networks for Missing Data

被引:4
|
作者
Likowski, Marcin Przewiez [1 ]
Smieja, Marek [1 ]
Struski, Lukasz [1 ]
Tabor, Jacek [1 ]
机构
[1] Jagiellonian Univ, Fac Math & Comp Sci, 6 Lojasiewicza St, PL-30348 Krakow, Poland
关键词
CHAINED EQUATIONS; IMPUTATION; MACHINE; VALUES;
D O I
10.1109/WACV51458.2022.00297
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Processing of missing data by modern neural networks, such as CNNs, remains a fundamental, yet unsolved challenge, which naturally arises in many practical applications, like image inpainting or autonomous vehicles and robots. While imputation-based techniques are still one of the most popular solutions, they frequently introduce unreliable information to the data and do not take into account the uncertainty of estimation, which may be destructive for a machine learning model. In this paper, we present MisConv, a general mechanism, for adapting various CNN architectures to process incomplete images. By modeling the distribution of missing values by the Mixture of Factor Analyzers, we cover the spectrum of possible replacements and find an analytical formula for the expected value of convolution operator applied to the incomplete image. The whole framework is realized by matrix operations, which makes MisConv extremely efficient in practice. Experiments performed on various image processing tasks demonstrate that MisConv achieves superior or comparable performance to the state-of-the-art methods.
引用
收藏
页码:2917 / 2926
页数:10
相关论文
共 50 条
  • [31] Convolutional Recurrent Neural Networks for Hyperspectral Data Classification
    Wu, Hao
    Prasad, Saurabh
    REMOTE SENSING, 2017, 9 (03)
  • [32] On the Impact of Imbalanced Data in Convolutional Neural Networks Performance
    Pulgar, Francisco J.
    Rivera, Antonio J.
    Charte, Francisco
    del Jesus, Maria J.
    HYBRID ARTIFICIAL INTELLIGENT SYSTEMS, HAIS 2017, 2017, 10334 : 220 - 232
  • [33] Data demodulation using convolutional neural networks for holographic data storage
    Katano, Yutaro
    Muroi, Tetsuhiko
    Kinoshita, Nobuhiro
    Ishii, Norihiko
    Hayashi, Naoto
    JAPANESE JOURNAL OF APPLIED PHYSICS, 2018, 57 (09)
  • [34] Inference of Missing PV Monitoring Data using Neural Networks
    Koubli, Eleni
    Palmer, Diane
    Betts, Tom
    Rowley, Paul
    Gottschalg, Ralph
    2016 IEEE 43RD PHOTOVOLTAIC SPECIALISTS CONFERENCE (PVSC), 2016, : 3436 - 3440
  • [35] Treatment of missing data using neural networks and genetic algorithms
    Abdella, M
    Marwala, T
    PROCEEDINGS OF THE INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), VOLS 1-5, 2005, : 598 - 603
  • [36] Estimation of Missing Data of Showcase Using Artificial Neural Networks
    Sakurai, Daiji
    Fukuyama, Yoshikazu
    Santana, Adamo
    Kawamura, Yu
    Murakami, Kenya
    Iizaka, Tatsuya
    Matsui, Tetsuro
    2017 IEEE 10TH INTERNATIONAL WORKSHOP ON COMPUTATIONAL INTELLIGENCE AND APPLICATIONS (IWCIA), 2017, : 15 - 18
  • [37] Filling missing suspended sediment data by artificial neural networks
    Cigizoglu, HK
    COMPUTATIONAL METHODS IN WATER RESOURCES, VOLS 1 AND 2, PROCEEDINGS, 2002, 47 : 1645 - 1652
  • [38] Infilling Missing Daily Evapotranspiration Data Using Neural Networks
    Abudu, Shalamu
    Bawazir, A. Salim
    King, J. Phillip
    JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING, 2010, 136 (05) : 317 - 325
  • [39] Using artificial neural networks to estimate missing rainfall data
    Kuligowski, RJ
    Barros, AP
    JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION, 1998, 34 (06): : 1437 - 1447
  • [40] Convolutional neural networks
    Alexander Derry
    Martin Krzywinski
    Naomi Altman
    Nature Methods, 2023, 20 : 1269 - 1270