Sequential Attention and Saccadic Grouping for Object Discrimination Using Reinforcement Learning

Lucas Paletta, Christin Seifert, Gerald Fritz

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


    An important issue in sequential object recognition is to de ne a strategy for saccadic access of visual information, and the representation of the features under observation. The original contribution of this work is to embed the saccadic decision procedure in a cascaded recognition process where visual evidence is probed exclusively at salient image locations. In a rst processing stage, salient image locations are determined from the local entropy in object discrimination. Local information in terms of code book vector responses contribute to the recognition state in the MDP. A reinforcement learner performs then trial and error search on useful actions towards salient locations within a neighborhood, receiving reward from entropy decreases. The method is evaluated in experiments on object recognition using the COIL-20 database, proving the method being computationally feasible and providing high recognition rates.
    Original languageEnglish
    Title of host publicationProceedings of the 1st Austrian Cognitive Vision Workshop 2005
    EditorsM. Zillich, M. Vincze
    PublisherAustrian Computer Society
    Number of pages8
    ISBN (Print)3-85403-186-6
    Publication statusPublished - 1 Jan 2005
    Event1st Austrian Cognitive Vision Workshop, ACVW 2005 - Zell an der Pram, Austria
    Duration: 31 Jan 200531 Jan 2005
    Conference number: 1


    Conference1st Austrian Cognitive Vision Workshop, ACVW 2005
    Abbreviated titleACVW
    CityZell an der Pram


    Dive into the research topics of 'Sequential Attention and Saccadic Grouping for Object Discrimination Using Reinforcement Learning'. Together they form a unique fingerprint.

    Cite this