Empirical analysis of cascade deformable models for multi-view face detection

J. Orozco, B. Martinez, M. Pantic

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    93 Downloads (Pure)

    Abstract

    In this paper, we present a face detector based on Cascade Deformable Part Models (CDPM) [1]. Our model is learnt from partially labelled images using Latent Support Vector Machines (LSVM). Recently Zhu et al. [2] proposed a Tree StructureModel for multi-view face detection trained with facial landmark labels, which resulted on a complex and suboptimal system for face detection. Instead, we adopt CDPMs enhanced with a data-mining procedure to enrich models during the LSVM training. Furthermore, a post-optimization procedure is derived to improve the performance of the CDPMs. Experimental results show that the proposed model can deal with highly expressive and partially occluded faces while outperforming the state-of-the-art face detectors by a large margin on challenging benchmarks such as the FDDB [3] and the AFLW [4] databases.
    Original languageEnglish
    Title of host publicationProceedings of the IEEE International Conference on Image Processing, ICIP 2013
    PublisherIEEE Computer Society
    Number of pages5
    Publication statusPublished - Sep 2013
    EventIEEE International Conference on Image Processing, ICIP 2013 - Melbourne, Australia
    Duration: 15 Sep 201318 Sep 2013

    Conference

    ConferenceIEEE International Conference on Image Processing, ICIP 2013
    Abbreviated titleICIP
    CountryAustralia
    CityMelbourne
    Period15/09/1318/09/13

    Keywords

    • HMI-HF: Human Factors
    • FDDB database
    • AFLW database
    • Multi-view face detection
    • Cascade deformable models

    Fingerprint Dive into the research topics of 'Empirical analysis of cascade deformable models for multi-view face detection'. Together they form a unique fingerprint.

    Cite this