Review finds unsupervised generative models show promise for neuroimaging anomaly detection

A systematic scoping review published on arXiv examines the application of unsupervised deep generative models for anomaly detection in neuroimaging. The review, which analyzed 33 studies from January 2018 to December 2025, found that these models, including autoencoders, VAEs, GANs, and diffusion models, show promise in identifying brain abnormalities without requiring detailed annotations. However, challenges such as methodological heterogeneity and limited external validation persist, though emerging techniques aim to improve robustness and clinical relevance. AI

IMPACT This review highlights the potential of AI in medical diagnostics, particularly for brain abnormalities where annotated data is scarce.

RANK_REASON The item is a systematic scoping review published on arXiv, detailing research methodology and findings. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• autoencoder
• CatalyzeX
• computed tomography
• DagsHub
• diffusion-based models
• generative adversarial network
• Gotit.pub
• Hugging Face
• magnetic resonance imaging
• PRISMA-ScR
• Variational Autoencoders
• Youwan MAHE

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →