PixelSmile: Toward Fine-Grained Facial Expression Editing

2026-03-26

AI Summary

This work addresses the challenge of fine-grained facial expression editing, which is hindered by semantic overlap among expressions and the difficulty of achieving precise, continuous, and identity-preserving control. To this end, the authors introduce the first Facial Expression Editing (FFE) dataset with continuous affective annotations, along with a comprehensive evaluation benchmark termed FFE-Bench. They further propose PixelSmile, a diffusion-based framework that disentangles expression semantics through fully symmetric joint training, integrating intensity supervision and contrastive learning to enable linearly interpolable expression control within the text latent space. PixelSmile is the first method to support smooth, fine-grained, and continuously controllable expression editing, significantly outperforming existing approaches in structural clarity, editing accuracy, linearity of control, and identity preservation.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Affective Computing

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Fine-grained facial expression editing has long been limited by intrinsic semantic overlap. To address this, we construct the Flex Facial Expression (FFE) dataset with continuous affective annotations and establish FFE-Bench to evaluate structural confusion, editing accuracy, linear controllability, and the trade-off between expression editing and identity preservation. We propose PixelSmile, a diffusion framework that disentangles expression semantics via fully symmetric joint training. PixelSmile combines intensity supervision with contrastive learning to produce stronger and more distinguishable expressions, achieving precise and stable linear expression control through textual latent interpolation. Extensive experiments demonstrate that PixelSmile achieves superior disentanglement and robust identity preservation, confirming its effectiveness for continuous, controllable, and fine-grained expression editing, while naturally supporting smooth expression blending.

Problem

• fine-grained facial expression editing
• semantic overlap
• identity preservation
• expression disentanglement
• continuous expression control

Innovation

• fine-grained facial expression editing
• diffusion model
• semantic disentanglement
• contrastive learning
• textual latent interpolation