Regularization approach is a key technique to improve the generalization ability of models. The current dropblock method removes information pixels on images by covering black pixels or patches of random noise. Such deletion is undesirable because it leads to information loss and inefficiency in the training period. Therefore, we propose to use better noise than vacancies for feature regularization, we use high-level feature map of the network as perturbations, and these high-level feature additions avoid losing useful information compared to vacancy additions, and this approach is easier to implement. It is experimentally demonstrated that our method has more boosting and generalization capability than dropblock.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Selfreg: Regularization Method via High-level Feature


    Contributors:


    Publication date :

    2021-10-20


    Size :

    1230362 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Enhancing feature tracking with gyro regularization

    Poling, B. / Lerman, G. | British Library Online Contents | 2016



    Part level transfer regularization for enhancing exemplar SVMs

    Aytar, Y. / Zisserman, A. | British Library Online Contents | 2015


    Semi-supervised sparse feature selection based on multi-view Laplacian regularization

    Shi, C. / Ruan, Q. / An, G. et al. | British Library Online Contents | 2015


    On the Length and Area Regularization for Multiphase Level Set Segmentation

    Bertelli, L. / Chandrasekaran, S. / Gibou, F. d. et al. | British Library Online Contents | 2010