Abstract:Background and ObjectiveThe skin lesion usually covers a small region of the dermoscopy image, and the lesions of different categories might own high similarities. Therefore, it is essential to design an elaborate network for accurate skin lesion classification, which can focus on semantically meaningful lesion parts. Although the Class Activation Mapping (CAM) shows good localization capability of highlighting the discriminative parts, it cannot be obtained in the forward propagation process.MethodsWe propose a Deep Attention Branch Network (DABN) model, which introduces the attention branches to expand the conventional Deep Convolutional Neural Networks (DCNN). The attention branch is designed to obtain the CAM in the training stage, which is then utilized as an attention map to make the network focus on discriminative parts of skin lesions. DABN is applicable to multiple DCNN structures and can be trained in an end-to-end manner. Moreover, a novel Entropy-guided Loss Weighting (ELW) strategy is designed to counter class imbalance influence in the skin lesion datasets.ResultsThe proposed method achieves an Average Precision (AP) of 0.719 on the ISIC-2016 dataset and an average area under the ROC curve (AUC) of 0.922 on the ISIC-2017 dataset. Compared with other state-of-the-art methods, our method obtains better performance without external data and ensemble learning. Moreover, extensive experiments demonstrate that it can be applied to multi-class classification tasks and improves mean sensitivity by more than 2.6% in different DCNN structures.ConclusionsThe proposed method can adaptively focus on the discriminative regions of dermoscopy images and allows for effective training when facing class imbalance, leading to the performance improvement of skin lesion classification, which could also be applied to other clinical applications.

Dense-Residual Attention Network for Skin Lesion Segmentation.

Attention-Guided Network with Densely Connected Convolution for Skin Lesion Segmentation

Skin Lesion Classification Using Densely Connected Convolutional Networks with Attention Residual Learning

ELA-Net: An Efficient Lightweight Attention Network for Skin Lesion Segmentation

EfficientSkinSegNet: A Lightweight ConvolutionalNeural Network for Accurate Skin LesionSegmentation

Deep attention branch networks for skin lesion classification

SEACU-Net: Attentive ConvLSTM U-Net with squeeze-and-excitation layer for skin lesion segmentation

Attention Residual Learning for Skin Lesion Classification

Skin lesion segmentation with attention-based SC-Conv U-Net and feature map distortion

[Image segmentation of skin lesions based on dense atrous spatial pyramid pooling and attention mechanism]

EAAC-Net: An Efficient Adaptive Attention and Convolution Fusion Network for Skin Lesion Segmentation

Channel Attention Separable Convolution Network for Skin Lesion Segmentation

AS-Net: Attention Synergy Network for skin lesion segmentation

Efficient and real-time skin lesion image segmentation using spatial-frequency information and channel convolutional networks

Automatic Skin Lesion Segmentation Using Deep Fully Convolutional Networks With Jaccard Distance

GA-Net: Ghost convolution adaptive fusion skin lesion segmentation network

HDS-Net: Achieving fine-grained skin lesion segmentation using hybrid encoding and dynamic sparse attention

RA-Net: Region-Aware Attention Network for Skin Lesion Segmentation

ACFNet: an Adaptive Context Fusion Network for Skin Lesion Segmentation

NCRNet: Neighborhood Context Refinement Network for skin lesion segmentation

Automatic skin lesion classification using a novel densely connected convolutional network integrated with an attention module