Patch vs. Global Image-Based Unsupervised Anomaly Detection in MR Brain Scans of Early Parkinsonian Patients

Verónica Muñoz-Ramírez (1, 2) , Nicolas Pinon (3) , Florence Forbes (2) , Carole Lartizien (3) , Michel Dojat (1)

1 GIN - [GIN] Grenoble Institut des Neurosciences
2 STATIFY - Modèles statistiques bayésiens et des valeurs extrêmes pour données structurées et de grande dimension
3 MYRIAD - Modeling & analysis for medical imaging and Diagnosis

Although neural networks have proven very successful in a number of medical image analysis applications, their use remains difficult when targeting subtle tasks such as the identification of barely visible brain lesions, especially given the lack of annotated datasets. Good candidate approaches are patch-based unsupervised pipelines which have both the advantage to increase the number of input data and to capture local and fine anomaly patterns distributed in the image, while potential inconveniences are the loss of global structural information. We illustrate this trade-off on Parkinson's disease (PD) anomaly detection comparing the performance of two anomaly detection models based on a spatial auto-encoder (AE) and an adaptation of a patch-fed siamese auto-encoder (SAE). On average, the SAE model performs better, showing that patches may indeed be advantageous.

Mots clés: Parkinson's disease; Anomaly detection; Patches; Siamese networks; Auto-encoders

Domaines: Intelligence artificielle [cs.AI]; Imagerie médicale; Neurosciences

Verónica Muñoz-Ramírez, Nicolas Pinon, Florence Forbes, Carole Lartizien, Michel Dojat. Patch vs. Global Image-Based Unsupervised Anomaly Detection in MR Brain Scans of Early Parkinsonian Patients. MLCN 2021 - 4th International Workshop in Machine Learning in Clinical Neuroimaging, Sep 2021, Strasbourg, France. pp.34-43, ⟨10.1007/978-3-030-87586-2_4⟩. ⟨hal-03397081⟩