Improving the Facial Expression Recognition and Its Interpretability via Generating Expression Pattern-map

作者：

Highlights：

• The EMR refines the expression attention mask by modeling the relationship among expression-related regions. The, EPMG generates a compact and discriminating embedding for expression recognition.

• Propose the concept of expression pattern-map which provides a unified visualization of expression features and improves the interpretability of facial expression recognition.

• Extensive experiments and empirical analysis are provided to demonstrate the effectiveness of the proposed method.

摘要

•Propose a method for facial expression recognition containing three modules: the Expression Feature Extractor (EFE), the Expression Mask Refiner (EMR), and the Expression Pattern-Map Generator (EPMG).•The EMR refines the expression attention mask by modeling the relationship among expression-related regions. The, EPMG generates a compact and discriminating embedding for expression recognition.•Propose the concept of expression pattern-map which provides a unified visualization of expression features and improves the interpretability of facial expression recognition.•Extensive experiments and empirical analysis are provided to demonstrate the effectiveness of the proposed method.

论文关键词：Facial expression recognition,Facial expression visualization,Expression pattern-map generator,Deep neural networks

论文评审过程：Received 13 November 2020, Revised 30 December 2021, Accepted 23 April 2022, Available online 2 May 2022, Version of Record 9 May 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108737