Fig 1.
Overall structure of EFFNet.
Fig 2.
Comparison of hair noise removal.
(A)is the original image and (B) is the image after hair noise removal.
Fig 3.
Image enhancement.
Fig 4.
(A) is the original image and (B) is the cropped image.
Fig 5.
MBConv layer and Fused-MBConv layer architecture.
Fig 6.
Modified EfficientNetV2-M model structure.
Fig 7.
Hierarchical bilinear pooling network architecture.
Fig 8.
Structure diagram of efficient channel attention mechanism.
Fig 9.
ECA-HBP structure diagram.
Fig 10.
Schematic diagram of random forests algorithm.
Table 1.
Distribution of different types of data in the dataset.
Fig 11.
Classification accuracy of different projection dimensions on the HAM10000 dataset.
Fig 12.
The confusion matrix of skin cancer classification model.
Table 2.
Evaluation indicators for each category of the model.
Table 3.
Ablation experiments.
Table 4.
Overall comparison of EFFNet with other state-of-the-art models on the test dataset.
Table 5.
Overall comparison of EFFNet with existing models on HAM10000 dataset.