In robotic applications often highly specific objects need to be recognized, e.g. industrial parts, for which methods can’t rely on the online availability of large labeled training data sets or pre-trained mod- els. This is especially valid for depth data, thus making it challenging for deep learning (DL) approaches. Therefore, this work analyzes the per- formance of various traditional (global or part-based) and DL features on a restricted depth data set, depending on the tasks complexity. While the sample size is small, we can conclude that pre-trained DL descriptors are the most descriptive but not by a statistically significant margin and therefore part-based descriptors are still a viable option for small but difficult 3D data sets.


    Access

    Download


    Export, share and cite



    Title :

    Applicability of Deep Learned vs Traditional Features for Depth Based Classification


    Contributors:
    Bracci, F (author) / Li, M (author) / Marton, Zoltan-Csaba (author) / Kossyk, Ingo (author)

    Conference:



    Publication date :

    2018-07-01



    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English