Skip to content
Book Open access

Machine-oriented Visual Media Quality Assessment

Nov 2026 · Proceedings of the 34th ACM International Conference on Multimedia · 0 citations · 40 references

Abstract

With the development of Embodied AI, machines have replaced humans as the main consumers of visual media, yet existing Image Quality Assessment (IQA) metrics remain focused on humans and overlook machine preference (a domain where quality depends on task utility rather than perceptual fidelity alone). To address this gap, we introduce the Machine-oriented Image Quality Assessment (MoIQA) Challenge, emphasizing simulation comprehension and real-world execution. The challenge comprises three tasks: 1) MoIQA-Absolute, primarily evaluating the consistency between IQA models and the performance of mainstream Vision-Language Models (VLMs) in simulation software; 2) MoIQA-Relative, measuring the consistency between IQA models and the variance of VLM inference after distortion, also in simulation; and 3) MoIQA-Real, focusing on whether IQA models match the results of the latest Vision-Language-Action Models (VLAs) in the Real-world. This challenge aims to advance reliable image understanding for machine preference and support robust Embodied AI applications. This grand challenge attracted 142 high-level teams around the world, and the relevant resources are open-sourced at: Homepage: https://m-vqa-challenge.github.io/M-VQA/ Dataset: https://huggingface.co/datasets/AIBench/M-VQA

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.