Journal: iScience
Article Title: Deep learning-enabled quantification of simultaneous PET/MRI for cell transplantation monitoring
doi: 10.1016/j.isci.2023.107083
Figure Lengend Snippet: Comparison, testing, and training of 3D CNN (A) Bar graph showing comparison of predictive performance of the 3D CNN, 2D CNN, and GBDT algorithms via analysis of root mean squared error (RMSE) and mean absolute error (MAE) (∗p < 0.05, ∗∗p < 0.05, the Student’s t test). (B) Bar graph showing comparison of Pearson correlation coefficient (PCC) values between 3D CNN, 2D CNN, and GBDT algorithms. (C) Loss graph of ten training rounds (iterations) comparing the 3D CNN and 2D CNN for both dataset 1 and dataset 2. (D) Intraclass correlation coefficient analysis of 3D CNN’s predictive accuracy on a range of images with 0 to 70μCi radioactivity (∗p = 0.001, ∗∗p = 0.001, the F test).
Article Snippet: Codes for K-mean ++, 2D CNN, 3D CNN and GBDT , This paper; Mendeley Data , https://doi.org/10.17632/995vjkhm68.1.
Techniques: Comparison, Radioactivity