Core Concepts
Customizable avatars with dynamic facial expressions improve user engagement in health applications.
Abstract
The content discusses the development of CADyFACE, a system using customizable avatars with dynamic facial expressions for health applications. It covers the importance of facial expressions in avatar-based stimuli, the proposed BeCoME-Net neural network for AU detection, and a feasibility study to evaluate construct validity.
Index:
Introduction to 3D avatars in health applications.
Importance of facial expressions in behavioral biomarker discovery.
Proposal of CADyFACE system and BeCoME-Net neural network.
Feasibility study design and tasks.
Key Highlights:
Avatars are effective tools for neuro-rehabilitation and therapeutic interventions.
CADyFACE aims to improve user engagement through customizable avatars with labeled expressions.
BeCoME-Net is proposed for AU detection using a novel correlation loss approach.
Feasibility study involves recognition and mimicry tasks to assess construct validity.
Stats
この研究は、National Science Foundationからの助成金によって支援されています。
BeCoME-Netは、AU検出のための新しいベータガイド相関損失を提案しています。