Unified model improves autism recognition from brain and face data
UniAR: A Unified Framework for Autism Recognition Enhanced by Multi-View Prompt Learning
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Autism is hard to diagnose early because there isn’t much written data available for training computer models. The authors created UniAR, a system that uses a big model to generate meaningful diagnostic descriptions from images of brains and faces. This helps the system understand the connection between what it sees and medical language, improving diagnosis accuracy. Tests show UniAR works better than previous methods on multiple datasets with brain and facial data.
What this means in practice
- •For medical imaging specialists: Use UniAR’s combined semantic descriptions and visual analysis to improve autism screening accuracy from brain MRI scans.
- •For clinical diagnostic teams: Apply UniAR’s multimodal approach to enhance autism detection from facial expressions when textual diagnostic data is scarce.
Authors
Lei Xin, Zeheng Wang, Jiayin Zhu, Shihong Huang, Fanhu Zeng, Changjiang Jiang, Dengbo He, Yutao Yue, Zhenglun Kong
Abstract
Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder for which early and accurate diagnosis is critical to improving long-term developmental outcomes. However, existing ASD recognition methods are often constrained by the scarcity of diagnostic text data, forcing them to rely mainly on visual analysis and limiting their ability to model clinically meaningful semantic reasoning. To address this challenge, we propose UniAR, a unified framework enhanced by multi-granularity prompt learning for robust ASD recognition under heterogeneous data variations. Specifically, UniAR leverages a large multimodal model to generate hierarchical diagnostic descriptions at the word, phrase, and sentence levels, compensating for the lack of paired clinical reports. To align the generated semantics with visual evidence, we further design a Mixture-of-Experts-based Multi-Scale Alignment Module, which dynamically matches vector-quantized visual prototypes with semantic representations at corresponding granularities. Extensive experiments on four benchmarks covering brain MRI and facial expression scenarios show that UniAR consistently outperforms existing state-of-the-art methods, achieving average accuracies of 75.9\% on MRI benchmarks and 91.6\% on facial benchmarks, while improving average Accuracy on MRI benchmarks by 1.5 percentage points and average Accuracy on facial benchmarks by 1.2 percentage points over baselines. These results demonstrate that UniAR offers a robust and interpretable framework for ASD screening under semantic scarcity.