Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Chi-en Amy Tai, Elizabeth Janes, Chris Czarnecki, Alexander Wong

TL;DR

This paper introduces a deep neural network design with DC-AC customized for skin cancer detection from skin lesion images and demonstrates that this approach with only 1.6 million parameters and 0.32 GFLOPs achieves better performance compared to traditional architecture designs.

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 compute 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. 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.

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

TL;DR

This paper introduces a deep neural network design with DC-AC customized for skin cancer detection from skin lesion images and demonstrates that this approach with only 1.6 million parameters and 0.32 GFLOPs achieves better performance compared to traditional architecture designs.

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 compute 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. 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.
Paper Structure (9 sections, 5 figures, 1 table)

This paper contains 9 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Works
  3. Materials and Methods
  4. Data Preparation
  5. Double-Condensing Attention Condenser Architecture Design
  6. Training Strategy
  7. Results
  8. Discussion
  9. Conclusions

Figures (5)

  • 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 wongFasterAttentionWhat2022. The numbers each layer is annotated with correspond to the depth dimension of the layer.
  • Figure 5: Sample skin lesions where the DC-AC design gave a correct prediction but the Cancer-Net SCa suite was incorrect.