Fully Automated Bone Age Assessment on Large-Scale Hand X-Ray Dataset

Algorithm 1.

Hand image segmentation.

Input:

: initial labeled training data, composed of samples.

: initial unlabeled training data, composed of samples.

Output:

: trained committee of U-Nets.

Repeat:

1. Train with the loss function in EQ2 on

2. Calculating unlabeled data’s uncertainty between different U-Nets and select the data with the largest uncertainty

3. Annotate the selected data by Oracle and add them into

Until: hand segmentation is satisfied on .