Bidirectional Cross-Scale Feature Fusion for Long Video Micro-Expression 3D Spotting Network

EasyChair Preprint 9818

Abstract

Psycho-cognitive computing is an integral part of intelligent human-computer interaction technology, which has received extensive attention in recent years. The research of micro-expressions can reflect the depth and breadth of mental cognitive computing. Micro-expression is a spontaneous, short-lived, and inadvertent facial expression. The research on micro-expression is of great significance in sentiment analysis, criminal investigation, and psychology. Micro-expression spotting refers to locating sequences of micro-expressions in a long video. Micro-expression detection is a crucial step in the micro-expression analysis. Based on the I3D backbone network of long video optical flow extraction and original video feature extraction, this paper extracts effective feature layers, performs downsampling, and uses the improved BiFPN module to fuse the extracted multi-scale feature layers. The final classification-regression network discriminates the detected expressions and locates the temporal boundaries where the expressions occur. The experimental results show that the proposed method effectively improves the F1 score index. Compared with other deep learning methods, this method performs better on CAS(ME)$^2$ and SAMM.

Keyphrases: 3D convolutional neural, BiFPN, Macro- and micro-expression spotting, Psycho-cognitive computing, deep learning, networks, optical flow

@booklet{EasyChair:9818,
  author    = {Xiaosong He and Xiao Wu and Jun Peng and Qingxia Li and Xinkai Ma and Yuanmin He},
  title     = {Bidirectional Cross-Scale Feature Fusion for Long Video Micro-Expression 3D Spotting Network},
  howpublished = {EasyChair Preprint 9818},
  year      = {EasyChair, 2023}}