Double-Condensing Attention Condenser: Leveraging Attention in Deep Learning to Detect Skin Cancer from Skin Lesion Images

Abstract

Skin cancer is the most common type of cancer in the United States and is estimated to affect one in five Americans. Recent advances have demonstrated strong performance on skin cancer detection, as exemplified by state of the art performance in the SIIM-ISIC Melanoma Classification Challenge; however, these solutions leverage ensembles of complex deep neural architectures requiring immense storage and computation costs, and therefore may not be tractable. A recent movement for TinyML applications is integrating Double-Condensing Attention Condensers (DC-AC) into a self-attention neural network backbone architecture to allow for faster and more efficient computation. This paper explores leveraging an efficient self-attention structure to detect skin cancer in skin lesion images and introduces a deep neural network design with DC-AC customized for skin cancer detection from skin lesion images. We demonstrate that our approach with only 1.6 million parameters and 0.32 GFLOPs achieves better performance compared to traditional architecture designs as it obtains an area under the ROC curve of 0.90 on the public ISIC 2020 test set and 0.89 on the private ISIC test set, over 0.13 above the best Cancer-Net SCa network architecture design. The final model is publicly available as a part of a global open-source initiative dedicated to accelerating advancement in machine learning to aid clinicians in the fight against cancer. Future work of this research includes iterating on the design of the selected network architecture and refining the approach to generalize to other forms of cancer.

Keywords: deep learning; self-attention; skin cancer.

Figure 1

Example skin lesion images.

Figure 2

Sample skin lesions that look like the other class.

Figure 3

Sample images that include additional elements beyond the skin lesion.

Figure 4

DC-AC architecture with condenser layers (orange), embedding layers (yellow), and expansion layers (blue) comprising DC-AC modules [10]. The numbers each layer is annotated with correspond to the depth dimension of the layer. DC-AC modules leverage an additional condensation step before selective attention is applied, leading to a more computationally efficient feature representation.

Figure 5

Sample skin lesions where the DC-AC design gave a correct prediction but the Cancer-Net SCa suite was incorrect.