MAC’26: The 3rd Micro-Action Analysis Grand Challenge

Dan Guo, Xiaobai Li, Kun Li, Haoyu Chen, Guoying Zhao, Meng Wang

Keywords: Micro-Action Understanding, Multi-label Micro-Action Detection

TL;DR: Building on MAC 2024/2025 at ACM MM 2024/2025, MAC 2026 adds the new MABench and introduces MAR, MMAD, and FMAU to advance fine-grained whole-body micro-action understanding.

Abstract: The **Micro-Action Analysis Grand Challenge (MAC 2026)** builds on MAC 2024 and MAC 2025, with a specific focus on fine-grained micro-action understanding. This edition targets the automatic analysis of subtle whole-body behaviors using computer vision and machine learning. Micro-actions are spontaneous, often fleeting body movements that reflect genuine emotions and intentions, yet are difficult to recognize and distinguish due to their subtle appearance. MAC 2026 comprises three tasks: **Micro-Action Recognition (MAR)**, **Multi-label Micro-Action Detection (MMAD)**, and a newly introduced task on **Fine-grained Micro-Action Understanding (FMAU)**, based on the **MA-52**, **MMA-52**, and **MABench** datasets, which contain **52 micro-action categories** and **8 annotated body parts**. The goal is to promote innovative research on fine-grained whole-body micro-actions, establish strong benchmarks, and advance technologies for human behavior understanding and emotional state analysis.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 1