Skip to main content
. Author manuscript; available in PMC: 2023 Dec 1.
Published in final edited form as: Med Image Anal. 2023 Sep 14;90:102960. doi: 10.1016/j.media.2023.102960

Fig. 8.

Fig. 8.

Classification performance of the proposed method on BUSI-derived datasets with varied numbers of location labels. The numbers 25%, 50%, 75%, and 100% on the x-axis represent the ratio of training samples with location labels compared to the original BUSI dataset in which all samples have both classification and localization labels. Each point represents the mean of each metric at the given labeling ratio, and the error bar shows its 95% confidence interval. In this experiment, ResNet50 was used as the FEX, and the training image size was 256 × 256 pixels.