ML Anthology

EfficientARL: Improving Skin Cancer Diagnoses by Combining Lightweight Attention on EfficientNet

Miguel Nehmad Alche, Daniel G. Acevedo, Marta Mejail
ICCVW 2021 pp. 3347-3353
doi:10.1109/ICCVW54120.2021.00374 /iccvw/2021/alche2021iccvw-efficientarl/

Abstract

Melanoma is a very dangerous form of skin cancer. Early diagnosis is crucial to increase the chances of its cure. Based on this, computer vision algorithms can be used to analyze dermoscopic images of skin lesions and decide if these correspond to benign or malignant tumors. In this work we propose the adaptation of the attention residual learning designed for ResNets to the EfficientNet networks, and compare this mechanism with attention mechanisms that these networks already have. We maintain the efficiency of these networks since only one extra parameter per stage needs to be trained. We also test several preprocessing methods on the dataset improving the final performance.

ICCVW Semantic Scholar

Cite

Text

Alche et al. "EfficientARL: Improving Skin Cancer Diagnoses by Combining Lightweight Attention on EfficientNet." IEEE/CVF International Conference on Computer Vision Workshops, 2021. doi:10.1109/ICCVW54120.2021.00374

Markdown

[Alche et al. "EfficientARL: Improving Skin Cancer Diagnoses by Combining Lightweight Attention on EfficientNet." IEEE/CVF International Conference on Computer Vision Workshops, 2021.](https://mlanthology.org/iccvw/2021/alche2021iccvw-efficientarl/) doi:10.1109/ICCVW54120.2021.00374

BibTeX

@inproceedings{alche2021iccvw-efficientarl,
  title     = {{EfficientARL: Improving Skin Cancer Diagnoses by Combining Lightweight Attention on EfficientNet}},
  author    = {Alche, Miguel Nehmad and Acevedo, Daniel G. and Mejail, Marta},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2021},
  pages     = {3347-3353},
  doi       = {10.1109/ICCVW54120.2021.00374},
  url       = {https://mlanthology.org/iccvw/2021/alche2021iccvw-efficientarl/}
}