Facial expression recognition is a basic task in computer vision, which has wide applications, such as human interaction, image recognition, etc. Numerous frameworks have been proposed for better accuracy and speed. In this work the author aim to exploit the effectiveness of visual attention mechanism to this task and investigate classical convolutional neural networks models for large scale object classification and have applied two prominent models VGG-16(very deep neural networks) network and SENet(Squeeze-and-Excitation Networks) with dataset in Kaggle in the field of facial expressions classification to better explore their accuracy. Cause VGG16 do have some defaults, SE block in SENet is added in VGG16 for better accuracy. The experiment on FER2013 dataset validates the effectiveness of the method proposed as before. Our method provides a comprehensive analysis for exploiting the visual attention modules to facial expression recognition. Our proposed algorithm can be equipped into other popular backbone to achieve better performance.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Facial Expression Algorithm using Attention Mechanism


    Contributors:
    Zou, Yiyang (author)


    Publication date :

    2022-10-12


    Size :

    1294029 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Multiscale Feature Fusion Attention Lightweight Facial Expression Recognition

    Jinyuan Ni / Xinyue Zhang / Jianxun Zhang | DOAJ | 2022

    Free access

    Facial Expression Hallucination

    Su, Congyong / Huang, Li | IEEE | 2005


    Facial Expression Decomposition

    Wang, H. / Ahuja, N. / IEEE | British Library Conference Proceedings | 2003


    Facial expression decomposition

    Hongcheng Wang, / Ahuja, | IEEE | 2003


    Drowsiness Detection Using Facial Expression Features

    Nakatani, Hiroto / Ishida, Kenji / Ozaki, Noriyuki et al. | SAE Technical Papers | 2010