ВИКОРИСТАННЯ ПРОСТОРОВИХ ВАГОВИХ МАСОК З НЕЗБАЛАНСОВАНИМИ ДАТАСЕТАМИ ДЛЯ ЗАДАЧІ СЕГМЕНТАЦІЇ СУПУТНИКОВИХ ЗНІМКІВ

Abstract

This study addresses the issue of imbalanced datasets in satellite image segmentation tasks, where machine learning models often neglect minority classes in favor of majority ones. We propose using spatial weight masks for the loss function computation to take into account the reliability score of individual pixels. This approach enhances segmentation quality, significantly improving metrics for minority classes. Additionally, a dataset augmentation method using generative adversarial networks (GANs) is explored, showing slight improvements in recognizing less represented crop types in the dataset, and it is compared with the proposed method. The simultaneous usage of weighted masks and generative networks is investigated. Keywords: dataset quality assessment, imbalanced datasets, classification, segmentation, generative adversarial networks, training data generation.