Seo, Hyunseok Bassenne, Maxime Xing, Lei 2024-01-19T15:32:26Z 2024-01-19T15:32:26Z 2021-09-02 2021-02 0278-0062 https://pubs.kist.re.kr/handle/201004/117484 Deep learning is becoming an indispensable tool for various tasks in science and engineering. A critical step in constructing a reliable deep learning model is the selection of a loss function, which measures the discrepancy between the network prediction and the ground truth. While a variety of loss functions have been proposed in the literature, a truly optimal loss function that maximally utilizes the capacity of neural networks for deep learning-based decision-making has yet to be established. Here, we devise a generalized loss function with functional parameters determined adaptively during model training to provide a versatile framework for optimal neural network-based decision-making in small target segmentation. The method is showcased by more accurate detection and segmentation of lung and liver cancer tumors as compared with the current state-of-the-art. The proposed formalism opens new opportunities for numerous practical applications such as disease diagnosis, treatment planning, and prognosis. English IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC Closing the Gap Between Deep Neural Network Modeling and Biomedical Decision-Making Metrics in Segmentation via Adaptive Loss Functions Article 10.1109/TMI.2020.3031913 1 IEEE TRANSACTIONS ON MEDICAL IMAGING, v.40, no.2, pp.585 - 593 IEEE TRANSACTIONS ON MEDICAL IMAGING 40 2 585 593 scie scopus 000615044900012 2-s2.0-85100619608 Computer Science, Interdisciplinary Applications Engineering, Biomedical Engineering, Electrical & Electronic Imaging Science & Photographic Technology Radiology, Nuclear Medicine & Medical Imaging Computer Science Engineering Imaging Science & Photographic Technology Radiology, Nuclear Medicine & Medical Imaging Article Training Neural networks Measurement Adaptation models Decision making Deep learning Harmonic analysis Deep learning Decision making Loss function Machine learning Segmentation