MMA-DFER: MultiModal Adaptation of unimodal models for Dynamic Facial Expression Recognition in-the-wild

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chumachenko, Kateryna, Iosifidis, Alexandros, Gabbouj, Moncef
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914753806336000
author Chumachenko, Kateryna
Iosifidis, Alexandros
Gabbouj, Moncef
author_facet Chumachenko, Kateryna
Iosifidis, Alexandros
Gabbouj, Moncef
contents Dynamic Facial Expression Recognition (DFER) has received significant interest in the recent years dictated by its pivotal role in enabling empathic and human-compatible technologies. Achieving robustness towards in-the-wild data in DFER is particularly important for real-world applications. One of the directions aimed at improving such models is multimodal emotion recognition based on audio and video data. Multimodal learning in DFER increases the model capabilities by leveraging richer, complementary data representations. Within the field of multimodal DFER, recent methods have focused on exploiting advances of self-supervised learning (SSL) for pre-training of strong multimodal encoders. Another line of research has focused on adapting pre-trained static models for DFER. In this work, we propose a different perspective on the problem and investigate the advancement of multimodal DFER performance by adapting SSL-pre-trained disjoint unimodal encoders. We identify main challenges associated with this task, namely, intra-modality adaptation, cross-modal alignment, and temporal adaptation, and propose solutions to each of them. As a result, we demonstrate improvement over current state-of-the-art on two popular DFER benchmarks, namely DFEW and MFAW.
format Preprint
id arxiv_https___arxiv_org_abs_2404_09010
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MMA-DFER: MultiModal Adaptation of unimodal models for Dynamic Facial Expression Recognition in-the-wild
Chumachenko, Kateryna
Iosifidis, Alexandros
Gabbouj, Moncef
Computer Vision and Pattern Recognition
Machine Learning
Dynamic Facial Expression Recognition (DFER) has received significant interest in the recent years dictated by its pivotal role in enabling empathic and human-compatible technologies. Achieving robustness towards in-the-wild data in DFER is particularly important for real-world applications. One of the directions aimed at improving such models is multimodal emotion recognition based on audio and video data. Multimodal learning in DFER increases the model capabilities by leveraging richer, complementary data representations. Within the field of multimodal DFER, recent methods have focused on exploiting advances of self-supervised learning (SSL) for pre-training of strong multimodal encoders. Another line of research has focused on adapting pre-trained static models for DFER. In this work, we propose a different perspective on the problem and investigate the advancement of multimodal DFER performance by adapting SSL-pre-trained disjoint unimodal encoders. We identify main challenges associated with this task, namely, intra-modality adaptation, cross-modal alignment, and temporal adaptation, and propose solutions to each of them. As a result, we demonstrate improvement over current state-of-the-art on two popular DFER benchmarks, namely DFEW and MFAW.
title MMA-DFER: MultiModal Adaptation of unimodal models for Dynamic Facial Expression Recognition in-the-wild
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2404.09010