Fig. 2 | Scientific Data

Fig. 2

From: A comprehensive dataset of magnetic resonance enterography images with intestinal segment annotations

Fig. 2

Data collection procedure. (a) Initial Annotation: MRE data undergoes sparse annotations by doctors, algorithm processing for coarse segmentation, and final refinement by doctors (1.5–2.5 h/patient). (b) Model Training: The refined annotations are used to train a deep learning segmentation model. (c) Iterative Refinement: The model predicts segments on new data, which are corrected by doctors to further improve the model (0.5–1.0 h/patient).

Back to article page