Machine learning models reproducibility and validation for MR images recognition

Ekaterina Kondrateva1Maxim Sharaev1Evgeny Burnaev1Alexander Bernstein1Irina Samotaeva2, 3

1Skolkovo Institute of Science and Technology2Moscow Research and Clinical Center for Neuropsychiatry3Institute of Higher Nervous Activity and Neurophysiology

International Conference on Machine Vision 2019 (ICMV2019)


In the present work, we introduce a data processing and analysis pipeline, which ensures the reproducibility of machine learning models chosen for MR image recognition. The proposed pipeline is applied to solve the binary classification problems: epilepsy and depression diagnostics based on vectorized features from MR images. This model is then assessed in terms of classification performance, robustness and reliability of the results, including predictive accuracy on unseen data. The classification performance achieved with our approach compares favorably to ones reported in the literature, where usually no thorough model evaluation is performed.