NeuroCBIR: A Fast and Accurate Image Retrieval System for Whole-Brain and Region-Specific MRI
Abstract
Content-based image retrieval (CBIR) in neuroimaging enables the identification of structurally similar brain scans, supporting diagnosis, prognosis, and treatment planning; however, existing methods are often limited to small datasets, single brain regions, or coarse class labels, thereby restricting their clinical utility and generalizability. Here, we present NeuroCBIR, a framework for fast and flexible retrieval of both whole-brain and region-specific 3D T1w MRI scans. A total of 103 cortical and subcortical regions are extracted to enable both whole-brain and region-level queries. NeuroCBIR leverages latent representations learned by a variational autoencoder (VAE) combined with contrastive learning, producing scan-specific embeddings that capture anatomical patterns. These embeddings were evaluated for subject re-identification, zero-shot age prediction, and zero-shot multi-class pathology stratification. Re-identification performance was high across both whole-brain and brain-region levels (mean average precision across the top-5 retrieved images (mAP@5) >= 98.4%), with robust generalization across datasets and acquisition conditions. While NeuroCBIR is not trained for age prediction or pathology stratification, zero-shot evaluations for these two tasks demonstrate that the embeddings encode meaningful information for downstream tasks. Embedding extraction on a 4-core CPU required approximately 18.7 s per scan, whereas similarity search was effectively instantaneous (less than 0.01 s). NeuroCBIR is publicly available for brain MRI with more than 26,000 precomputed T1w MRI embeddings. It supports reproducible research, region-specific flexibility, and clinically meaningful personalized diagnostic support. The software is available at https://github.com/minnelab/NeuroCBIR.
Related
- An ER Retrieval-and-Release System Reveals Ligand-Independent Basal Signaling by Cell Surface TCRs
- Cortical high frequency oscillations reflect encoding and retrieval of specific word concepts
- Single-nucleus transcriptomics uncovers region-specific molecular features underlying selective aging vulnerability of cholinergic neurons
- Repeated VU0155069 Treatment Is Associated with Region- and Pathological Stage-Specific Dendritic Spine Remodeling in Female 3xTg-AD Mice
- Population Genomics of Almond ( Prunus dulcis ) Reveals Region‐Specific Selection and a Complex History of Domestication
BibTeX
@misc{nietodelamor2026neurocbir,
title = {{NeuroCBIR: A Fast and Accurate Image Retrieval System for Whole-Brain and Region-Specific MRI}},
author = {Felix Nieto-del-Amor and Jingru Fu and J. -Sebastian Muehlboeck and Eric Westman and Daniel Ferreira and Rodrigo Moreno},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.06502},
eprint = {2610.06502},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.06502},
}
RIS
TY - UNPB
TI - NeuroCBIR: A Fast and Accurate Image Retrieval System for Whole-Brain and Region-Specific MRI
AU - Felix Nieto-del-Amor
AU - Jingru Fu
AU - J. -Sebastian Muehlboeck
AU - Eric Westman
AU - Daniel Ferreira
AU - Rodrigo Moreno
PY - 2026
DA - 2026/10/05
PB - arXiv
DO - 10.48550/arxiv.2610.06502
UR - https://arxiv.org/abs/2610.06502
AB - Content-based image retrieval (CBIR) in neuroimaging enables the identification of structurally similar brain scans, supporting diagnosis, prognosis, and treatment planning; however, existing methods are often limited to small datasets, single brain regions, or coarse class labels, thereby restricting their clinical utility and generalizability. Here, we present NeuroCBIR, a framework for fast and flexible retrieval of both whole-brain and region-specific 3D T1w MRI scans. A total of 103 cortical and subcortical regions are extracted to enable both whole-brain and region-level queries. NeuroCBIR leverages latent representations learned by a variational autoencoder (VAE) combined with contrastive learning, producing scan-specific embeddings that capture anatomical patterns. These embeddings were evaluated for subject re-identification, zero-shot age prediction, and zero-shot multi-class pathology stratification. Re-identification performance was high across both whole-brain and brain-region levels (mean average precision across the top-5 retrieved images (mAP@5) >= 98.4%), with robust generalization across datasets and acquisition conditions. While NeuroCBIR is not trained for age prediction or pathology stratification, zero-shot evaluations for these two tasks demonstrate that the embeddings encode meaningful information for downstream tasks. Embedding extraction on a 4-core CPU required approximately 18.7 s per scan, whereas similarity search was effectively instantaneous (less than 0.01 s). NeuroCBIR is publicly available for brain MRI with more than 26,000 precomputed T1w MRI embeddings. It supports reproducible research, region-specific flexibility, and clinically meaningful personalized diagnostic support. The software is available at https://github.com/minnelab/NeuroCBIR.
ER -