Fig. 3: Classification outcomes by DL models.
From: Focal cortical dysplasia (type II) detection with multi-modal MRI and a deep-learning framework
a AUPR values obtained with DenseNet201 in classification of FCD-II subtypes (FCD-IIa and FCD-IIb) using Iacs of FLAIR and T1w MRI modes respectively (the first two images of the first and second rows, respectively), b The highest performance values (box plots) and corresponding AUPR values obtained with DenseNet201 model in the classification of FCD-IIa and FCD-IIb using 2D images of the individual sagittal plane of FLAIR (the last two representations of the first row) and T1w (the last two representation of the second row), where the whiskers range within 95% of the confidence interval. c The highest performance scores recorded with MobileNetV2 for Iacs (FLAIR) and the sagittal plane (T1w) in the classification of subtypes, along with the corresponding AUPR values, are shown.
