Patch2Loc: Learning to Localize Patches for Unsupervised Brain Lesion Detection
Not on the map. Why?
Abstract (arXiv)
Detecting brain lesions as abnormalities observed in magnetic resonance imaging (MRI) is essential for diagnosis and treatment. In the search of abnormalities, such as tumors and malformations, radiologists may benefit from computer-aided diagnostics that use computer vision systems trained with machine learning to segment normal tissue from abnormal brain tissue. While supervised learning methods require annotated lesions, we propose a new unsupervised approach (Patch2Loc) that learns from normal patches taken from structural MRI. We train a neural network model to map a patch back to its spatial location within a slice of the brain volume. During inference, abnormal patches are detected by the relatively higher error and/or variance of the location prediction. This generates a heatmap that can be integrated into pixel-wise methods to achieve finer-grained segmentation. We demonstrate the ability of our model to segment abnormal brain tissues by applying our approach to the detection of tumor tissues in MRI on T2-weighted images from BraTS2021 and MSLUB datasets and T1-weighted images from ATLAS and WMH datasets. We show that it outperforms the state-of-the art in unsupervised segmentation. The implementation for this work can be found on our hrefhttps://github.com/bakerhassan/Patch2LocGitHub page. This paper has been accepted at AISTATS 2026.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 23 of its 46 references that are in the map. First posted on arXiv 25 Jun 2025.
Builds on
23 of 46 references.
References · 23
- 2017Attention Is All You Needcited by 60,361
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2020Denoising Diffusion Probabilistic Modelscited by 21,076
- 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
- 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
- 2015Unsupervised Visual Representation Learning by Context Predictioncited by 1,066
- 2021The RSNA-ASNR-MICCAI BraTS 2021 Benchmark on Brain Tumor Segmentation and Radiogenomic Classificationcited by 518
- 2018Deep Autoencoding Models for Unsupervised Anomaly Segmentation in Brain MR Imagescited by 146
- 20203D Self-Supervised Methods for Medical Imagingcited by 134
- 2022On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networkscited by 120
- 2019Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities and Results of the WMH Segmentation Challengecited by 115
11 more
- 2022Fast Unsupervised Brain Anomaly Detection and Segmentation with Diffusion Modelscited by 97
- 2018Unsupervised Detection of Lesions in Brain MRI using constrained adversarial auto-encoderscited by 83
- 2020Unsupervised Lesion Detection via Image Restoration with a Normative Priorcited by 65
- 2018Context-encoding Variational Autoencoder for Unsupervised Anomaly Detectioncited by 64
- 2023Mask, Stitch, and Re-Sample: Enhancing Robustness and Generalizability in Anomaly Detection through Automatic Diffusion Modelscited by 50
- 2023Patched Diffusion Models for Unsupervised Anomaly Detection in Brain MRIcited by 45
- 2021Challenging Current Semi-Supervised Anomaly Segmentation Methods for Brain MRIcited by 23
- 2022Unsupervised Anomaly Localization with Structural Feature-Autoencoderscited by 23
- 2023Reversing the Abnormal: Pseudo-Healthy Generative Networks for Anomaly Detectioncited by 18
- 2023Guided Reconstruction with Conditioned Diffusion Models for Unsupervised Anomaly Detection in Brain MRIscited by 18
- 2023Modality Cycles with Masked Conditional Diffusion for Unsupervised Anomaly Segmentation in MRIcited by 12
Built on by
Citing papers per year: 2026: 1.
1 paper.
- 2026Normal Guidance is what Attention Needscited by 0
- 2026Normal Guidance is what Attention Needs26 May 2026