On Stereo Confidence Measures for Global Methods: Evaluation, New Model and Integration into Occupancy Grids

Martim Brandao, Ricardo Ferreira, Kenji Hashimoto, Atsuo Takanishi, Jose Santos-Victor

    Research output: Contribution to journalArticle

    5 Citations (Scopus)

    Abstract

    Stereo confidence measures are important functions for global reconstruction methods and some applications of stereo. In this article we evaluate and compare several models of confidence which are defined at the whole disparity range. We propose a new stereo confidence measure to which we call the Histogram Sensor Model (HSM), and show how it is one of the best performing functions overall. We also introduce, for parametric models, a systematic method for estimating their parameters which is shown to lead to better performance when compared to parameters as computed in previous literature. All models were evaluated when applied to two different cost functions at different window sizes and model parameters. Contrary to previous stereo confidence measure benchmark literature, we evaluate the models with criteria important not only to winner-take-all stereo, but also to global applications. To this end, we evaluate the models on a real-world application using a recent formulation of 3D reconstruction through occupancy grids which integrates stereo confidence at all disparities. We obtain and discuss our results on both indoors' and outdoors' publicly available datasets.

    Original languageEnglish
    Article number7112529
    Pages (from-to)116-128
    Number of pages13
    JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
    Volume38
    Issue number1
    DOIs
    Publication statusPublished - 2016 Jan 1

    Fingerprint

    Confidence Measure
    Evaluation Method
    Grid
    Confidence
    Evaluate
    Model
    Winner-take-all
    3D Reconstruction
    Real-world Applications
    Parametric Model
    Histogram
    Cost Function
    Cost functions
    Integrate
    Benchmark
    Sensor
    Formulation
    Sensors
    Range of data

    Keywords

    • 3D reconstruction
    • confidence
    • occupancy grids
    • stereo matching
    • Stereo vision
    • uncertainty

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Vision and Pattern Recognition
    • Software
    • Computational Theory and Mathematics
    • Applied Mathematics

    Cite this

    On Stereo Confidence Measures for Global Methods : Evaluation, New Model and Integration into Occupancy Grids. / Brandao, Martim; Ferreira, Ricardo; Hashimoto, Kenji; Takanishi, Atsuo; Santos-Victor, Jose.

    In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 38, No. 1, 7112529, 01.01.2016, p. 116-128.

    Research output: Contribution to journalArticle

    @article{d850d8f0b90c4c97a977ea148083c45d,
    title = "On Stereo Confidence Measures for Global Methods: Evaluation, New Model and Integration into Occupancy Grids",
    abstract = "Stereo confidence measures are important functions for global reconstruction methods and some applications of stereo. In this article we evaluate and compare several models of confidence which are defined at the whole disparity range. We propose a new stereo confidence measure to which we call the Histogram Sensor Model (HSM), and show how it is one of the best performing functions overall. We also introduce, for parametric models, a systematic method for estimating their parameters which is shown to lead to better performance when compared to parameters as computed in previous literature. All models were evaluated when applied to two different cost functions at different window sizes and model parameters. Contrary to previous stereo confidence measure benchmark literature, we evaluate the models with criteria important not only to winner-take-all stereo, but also to global applications. To this end, we evaluate the models on a real-world application using a recent formulation of 3D reconstruction through occupancy grids which integrates stereo confidence at all disparities. We obtain and discuss our results on both indoors' and outdoors' publicly available datasets.",
    keywords = "3D reconstruction, confidence, occupancy grids, stereo matching, Stereo vision, uncertainty",
    author = "Martim Brandao and Ricardo Ferreira and Kenji Hashimoto and Atsuo Takanishi and Jose Santos-Victor",
    year = "2016",
    month = "1",
    day = "1",
    doi = "10.1109/TPAMI.2015.2437381",
    language = "English",
    volume = "38",
    pages = "116--128",
    journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
    issn = "0162-8828",
    publisher = "IEEE Computer Society",
    number = "1",

    }

    TY - JOUR

    T1 - On Stereo Confidence Measures for Global Methods

    T2 - Evaluation, New Model and Integration into Occupancy Grids

    AU - Brandao, Martim

    AU - Ferreira, Ricardo

    AU - Hashimoto, Kenji

    AU - Takanishi, Atsuo

    AU - Santos-Victor, Jose

    PY - 2016/1/1

    Y1 - 2016/1/1

    N2 - Stereo confidence measures are important functions for global reconstruction methods and some applications of stereo. In this article we evaluate and compare several models of confidence which are defined at the whole disparity range. We propose a new stereo confidence measure to which we call the Histogram Sensor Model (HSM), and show how it is one of the best performing functions overall. We also introduce, for parametric models, a systematic method for estimating their parameters which is shown to lead to better performance when compared to parameters as computed in previous literature. All models were evaluated when applied to two different cost functions at different window sizes and model parameters. Contrary to previous stereo confidence measure benchmark literature, we evaluate the models with criteria important not only to winner-take-all stereo, but also to global applications. To this end, we evaluate the models on a real-world application using a recent formulation of 3D reconstruction through occupancy grids which integrates stereo confidence at all disparities. We obtain and discuss our results on both indoors' and outdoors' publicly available datasets.

    AB - Stereo confidence measures are important functions for global reconstruction methods and some applications of stereo. In this article we evaluate and compare several models of confidence which are defined at the whole disparity range. We propose a new stereo confidence measure to which we call the Histogram Sensor Model (HSM), and show how it is one of the best performing functions overall. We also introduce, for parametric models, a systematic method for estimating their parameters which is shown to lead to better performance when compared to parameters as computed in previous literature. All models were evaluated when applied to two different cost functions at different window sizes and model parameters. Contrary to previous stereo confidence measure benchmark literature, we evaluate the models with criteria important not only to winner-take-all stereo, but also to global applications. To this end, we evaluate the models on a real-world application using a recent formulation of 3D reconstruction through occupancy grids which integrates stereo confidence at all disparities. We obtain and discuss our results on both indoors' and outdoors' publicly available datasets.

    KW - 3D reconstruction

    KW - confidence

    KW - occupancy grids

    KW - stereo matching

    KW - Stereo vision

    KW - uncertainty

    UR - http://www.scopus.com/inward/record.url?scp=84961578459&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84961578459&partnerID=8YFLogxK

    U2 - 10.1109/TPAMI.2015.2437381

    DO - 10.1109/TPAMI.2015.2437381

    M3 - Article

    AN - SCOPUS:84961578459

    VL - 38

    SP - 116

    EP - 128

    JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

    JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

    SN - 0162-8828

    IS - 1

    M1 - 7112529

    ER -