https://github.com/rajsoni03/deme-net

A CNN architecture to Diagnosis of Dermatology Melanoma Skin Cancer using parallel convolution.
https://github.com/rajsoni03/deme-net

classification cnn dermatology keras melanoma parallel-algorithm skin-cancer tendorflow

Last synced: 11 months ago
JSON representation

A CNN architecture to Diagnosis of Dermatology Melanoma Skin Cancer using parallel convolution.

• Host: GitHub
• URL: https://github.com/rajsoni03/deme-net
• Owner: Rajsoni03
• License: mit
• Created: 2021-06-12T11:10:06.000Z (over 4 years ago)
• Default Branch: main
• Last Pushed: 2021-06-12T15:01:28.000Z (over 4 years ago)
• Last Synced: 2025-01-22T10:31:43.772Z (about 1 year ago)
• Topics: classification, cnn, dermatology, keras, melanoma, parallel-algorithm, skin-cancer, tendorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 1.27 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# DeMe-Net

### A CNN architecture to Diagnosis of Dermatology Melanoma Skin Cancer using parallel convolution.

### The DeMe Net architecture is achieves 98.33% accuracy on testing data and 99.85% on the traininge detaset.

### We used the HAM10000 dataset from Kaggle, which was published by ISIC for the 2018 ML Challenge.

![Screenshot-1](images/data_img.png)

### The DeMe net architecture used 3x3, 5x5, 7x7, and 11x11 parallel conv2d layar for distributed computing. 

![Screenshot-2](images/DeMe_Net.png)

### We have used accuracy, precision, recall and f1-score as performance metrics to test the performance of the model.

![Screenshot-3](images/confusion_matrix.png)

![Screenshot-4](images/classification_report.png)