Fig. 3 | Scientific Reports

Fig. 3

From: Deep-learning model accurately classifies multi-label lung ultrasound findings, enhancing diagnostic accuracy and inter-reader agreement

Fig. 3

Basic feature extractor structure of EfficientNet-B0. The input data is converted from DICOM to PNG through a hand-made algorithm that goes through the model. The model combines EfficientNet-B0 with fully connected layers consisting of 4 nodes, fitting the purpose of classification. Afterward, the prediction scores are calculated using a sigmoid function and Binary Cross Entropy to enable multi-label classification.

Back to article page