Unet-boosted classifier – мультизадачная архитектура для малых выборок на примере классификации МРТ снимков головного мозга

Информатика и автоматизация (Труды СПИИРАН). 2024. Т. 23. № 4. С. 1022–1046.

Sobyanin K., Kulikova S.

The problem of training deep neural networks on small samples is especially relevant for medical problems. The paper examines the impact of pixel-wise marking of significant objects in the image, over the true class label, on the quality of the classification. To achieve better classification results on small samples, we propose a multitasking architecture -- Unet-boosted classifier (UBC), that is trained simultaneously to solve classification and semantic segmentation problems. As the exploratory dataset, MRI images of patients with benign glioma and glioblastoma taken from the BRaTS 2019 data set are used. One horizontal slice of the MRI image containing a glioma is considered as the input (a total of 380 frames in the training set), and the probability of glioblastoma -- as the output. Resnet34 was chosen as the baseline, trained without augmentations with a loss function based on cross entropy. As an alternative solution UBC-resnet34 is used -- the same resnet34, boosted by a decoder built on the U-Net principle and predicting the pixels with glioma. The smoothed Sorensen–Dice coefficient (DiceLoss) is used as an decoder loss function. Results on the test sample: accuracy for the baseline reached 0.71, for the proposed model -- 0.81, the Dice score -- 0.77. Thus, a deep model can be well trained even on a small data set, using the proposed architecture, provided that marking of the affected tissues in the form of a semantic mask are provided.

Keywords: deep learning, глубокое обучение, Semantic segmentation, классификация изображений, image classification, семантическая сегментация, Small datasets, multi-task architecture, cerebral pathology, tumor diagnosis, малый набор данных, мультизадачная архитектура, церебральная патология, диагностика опухоли

Publication based on the results of: Разработка автоматических подходов для определения этиологии криптогенного инсульта с целью профилактики вторичных острых нарушений мозгового кровообращения (2023)