In-batch Relational Features Enhance Precision in An Unsupervised Medical Anomaly Detection Task
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908868719673344 |
|---|---|
| author | Githinji, P. Bilha Yuan, Xi Gul, Ijaz Zhang, Lian Xu, Jinhao Chen, Zhenglin Qin, Peiwu Yu, Dongmei |
| author_facet | Githinji, P. Bilha Yuan, Xi Gul, Ijaz Zhang, Lian Xu, Jinhao Chen, Zhenglin Qin, Peiwu Yu, Dongmei |
| contents | Confounding pathology with normal anatomical variation remains a significant challenge in unsupervised medical-image anomaly detection, resulting in numerous false positives. To enhance integration of healthy variation, we augment the latent representation of a CNN autoencoder with contextual similarities within a normal cohort through batch-wise hypergraph estimation and a shared-weights graph convolution layer, producing a population-aware embedding. On a heterogeneous brain-tumor dataset of 2D MRI scans, the method improves separability between healthy and pathological samples, achieving an AUC-ROC of 0.90 (95% CI 0.84-0.95, 5.7% absolute gain), and a 16% absolute improvement in average precision (0.78 AP, 95% CI 0.66-0.89), thereby lowering false-positive rates. Moreover, both anomaly detection and downstream tumor versus no-tumor classification performance improve with the size of the mini-batch context captured in the augmented representation, suggesting a tunable lever for integrating healthy variation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_05534 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | In-batch Relational Features Enhance Precision in An Unsupervised Medical Anomaly Detection Task Githinji, P. Bilha Yuan, Xi Gul, Ijaz Zhang, Lian Xu, Jinhao Chen, Zhenglin Qin, Peiwu Yu, Dongmei Quantitative Methods Image and Video Processing Confounding pathology with normal anatomical variation remains a significant challenge in unsupervised medical-image anomaly detection, resulting in numerous false positives. To enhance integration of healthy variation, we augment the latent representation of a CNN autoencoder with contextual similarities within a normal cohort through batch-wise hypergraph estimation and a shared-weights graph convolution layer, producing a population-aware embedding. On a heterogeneous brain-tumor dataset of 2D MRI scans, the method improves separability between healthy and pathological samples, achieving an AUC-ROC of 0.90 (95% CI 0.84-0.95, 5.7% absolute gain), and a 16% absolute improvement in average precision (0.78 AP, 95% CI 0.66-0.89), thereby lowering false-positive rates. Moreover, both anomaly detection and downstream tumor versus no-tumor classification performance improve with the size of the mini-batch context captured in the augmented representation, suggesting a tunable lever for integrating healthy variation. |
| title | In-batch Relational Features Enhance Precision in An Unsupervised Medical Anomaly Detection Task |
| topic | Quantitative Methods Image and Video Processing |
| url | https://arxiv.org/abs/2603.05534 |