https://github.com/eric11eca/melanoma-classification
https://github.com/eric11eca/melanoma-classification
Last synced: over 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/eric11eca/melanoma-classification
- Owner: eric11eca
- License: mit
- Created: 2020-06-23T10:25:49.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2020-07-22T18:31:53.000Z (almost 6 years ago)
- Last Synced: 2025-01-22T06:48:20.525Z (over 1 year ago)
- Language: Jupyter Notebook
- Size: 199 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# SIIM-ISIC Melanoma Classification
This repository contains scripts for the Kaggle competition for SIIM-ISIC Melanoma Classification[https://www.kaggle.com/c/siim-isic-melanoma-classification/overview]. The model is required to identify melanoma in images of skin lesions. In particular, images within the same patient will be classified as likely to represent a melanoma or not.
## Data
Bothe train and test data for the competition can be directly downloaded from the competition page or via Kaggle command line API, which can be installed using pip.
## Baseline Model
We selected the pretrained ResNet as the baseline model and plan to build a more roboust, accurate, and efficient model from there. Currently, only vision based deep neural networks are considered for the task. However, in the future, other machine learning models as well as other learning methods should be considered to aggregate both the visual features from the images and the contextual futures from the patients.