Fig. 6

Computational complexities of the proposed framework and some common methods for breast lesion classification for IHR data: (a) Total number of model parameters, (b) FLOPS. Note: P denotes pre-trained.

Computational complexities of the proposed framework and some common methods for breast lesion classification for IHR data: (a) Total number of model parameters, (b) FLOPS. Note: P denotes pre-trained.