Brain MRI Bilateral Symmetry Dataset

This dataset is the frozen data snapshot used for research on multi-descriptor bilateral spatial and statistical texture asymmetry for brain MRI abnormality identification. The dataset contains 6,106 MRI hemisphere images organised into 3,053 bilateral left-right image pairs, comprising 1,608 normal pairs and 1,445 abnormal pairs. Images are organised into four directories: NL (normal-left), NR (normal-right), YL (abnormal-left), and YR (abnormal-right). Left-right pairing is based on matching filename stems within each diagnostic class. Filename stems should not be interpreted as independently verified patient identifiers; consequently, the associated research evaluates pair-level rather than independently verified patient-level generalisation. This Zenodo release preserves the frozen image collection used for Paper 5 of the Symmetry Brain Research programme. A SHA-256 manifest covering all 6,106 images and a SHA-256 checksum for the complete dataset archive are provided to support data integrity verification and reproducibility. The underlying collection is derived from the “Brain Cancer MRI Images, Lobes Symmetry Dataset” originally released on Kaggle (DOI: 10.34740/KAGGLE/DSV/4710773). This dataset is intended for research and reproducibility in medical image analysis, bilateral brain symmetry, texture analysis, computer vision, and machine learning. It is not intended for direct clinical diagnosis. Database licence: Open Data Commons Open Database License (ODbL) 1.0. Database contents: Database Contents License (DbCL) 1.0.