Gamma: Toward Generic Image Assessment with Mixture of Assessment Experts

2025-03-09

AI Summary

Existing image assessment methods are often confined to specific subtasks or scenarios, whereas unified models struggle to balance performance and generalizability. To address this, we propose Gamma, a general-purpose image assessment model capable of joint quality and aesthetic evaluation across diverse domains—including natural images and AI-generated content (AIGC). To mitigate annotation bias from heterogeneous data sources and overcome challenges in scene-specific modeling, we introduce two novel components: the Mixture of Assessment Experts (MoAE) module and the Scene-Difference Prompting (SDP) strategy. These leverage a shared-adaptive expert architecture, dynamic routing, and a multi-dataset joint training framework. Extensive experiments demonstrate that Gamma consistently outperforms existing unified models across 12 benchmark datasets and 6 distinct assessment scenarios, achieving superior coverage and accuracy. The implementation is publicly available.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Mixture of Experts (MoE)
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Image assessment aims to evaluate the quality and aesthetics of images and has been applied across various scenarios, such as natural and AIGC scenes. Existing methods mostly address these sub-tasks or scenes individually. While some works attempt to develop unified image assessment models, they have struggled to achieve satisfactory performance or cover a broad spectrum of assessment scenarios. In this paper, we present extbf{Gamma}, a extbf{G}eneric im extbf{A}ge assess extbf{M}ent model using extbf{M}ixture of extbf{A}ssessment Experts, which can effectively assess images from diverse scenes through mixed-dataset training. Achieving unified training in image assessment presents significant challenges due to annotation biases across different datasets. To address this issue, we first propose a Mixture of Assessment Experts (MoAE) module, which employs shared and adaptive experts to dynamically learn common and specific knowledge for different datasets, respectively. In addition, we introduce a Scene-based Differential Prompt (SDP) strategy, which uses scene-specific prompts to provide prior knowledge and guidance during the learning process, further boosting adaptation for various scenes. Our Gamma model is trained and evaluated on 12 datasets spanning 6 image assessment scenarios. Extensive experiments show that our unified Gamma outperforms other state-of-the-art mixed-training methods by significant margins while covering more scenes. Code: https://github.com/zht8506/Gamma.

Problem

• Develops a unified image assessment model for diverse scenes.
• Addresses annotation biases across different datasets effectively.
• Enhances performance and coverage of image assessment scenarios.

Innovation

• Mixture of Assessment Experts (MoAE) module
• Scene-based Differential Prompt (SDP) strategy
• Mixed-dataset training for diverse scenes