Comparative experiment method (IMAGE)
Caption
A comparative experiment has been designed to validate our hypothesis as shown in the figure below. We used lung cancer and ABUS datasets to pre-train our previously proposed architecture. We focused on learning the common features of the two diseases and the empirical parameters obtained from the corresponding detection tasks. Subsequently, we obtained two pretrained models with two different datasets: pretrained models A and B. We switched the pretraining datasets for the two tasks to the same ABUS training set and continued to train the two different pretrained models. Following the training, we obtained Model D with lung cancer detection experience and Model C with only ABUS breast cancer detection experience. Finally, we tested the performances of models C and D using the same test set to validate our ideas.
Credit
Beijing Zhongke Journal Publising Co. Ltd.
Usage Restrictions
Credit must be given to the creator.
License
CC BY