https://github.com/manoharvit/clinical-skin-lesion-diagnosis

Evaluated custom CNN and VGG models to classify seven distinct categories of skin lesions for addressing the challenge of data imbalance in medical datasets. Employed advanced machine learning techniques, comparative analysis of architectures, to enhance diagnostic accuracy in dermatology, contributing to early skin cancer detection.
https://github.com/manoharvit/clinical-skin-lesion-diagnosis

Last synced: about 2 months ago
JSON representation

Evaluated custom CNN and VGG models to classify seven distinct categories of skin lesions for addressing the challenge of data imbalance in medical datasets. Employed advanced machine learning techniques, comparative analysis of architectures, to enhance diagnostic accuracy in dermatology, contributing to early skin cancer detection.

• Host: GitHub
• URL: https://github.com/manoharvit/clinical-skin-lesion-diagnosis
• Owner: ManoharVit
• Created: 2024-01-12T18:52:51.000Z (12 months ago)
• Default Branch: main
• Last Pushed: 2024-01-12T19:03:16.000Z (12 months ago)
• Last Synced: 2024-01-13T10:24:52.181Z (12 months ago)
• Language: Jupyter Notebook
• Size: 5.27 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Clinical-Skin-Lesion-Diagnosis

## Overview

This project focuses on the classification of dermatoscopic images of skin lesions into seven categories, using Deep Learning and Convolutional Neural Networks (CNNs). It addresses the critical issue of data imbalance in medical datasets using the "HAM10000" dataset, which contains diverse skin lesion images.

## Key Features

Implementation of a custom CNN model and a VGG-inspired model for medical image analysis.

Addressing class representation gaps in healthcare datasets with an emphasis on reducing class overrepresentation.

Comparative analysis of different CNN architectures to determine the most effective method for skin cancer diagnosis.

## Technologies

Deep Learning, Convolutional Neural Networks

Data Analysis and Visualization

Python, TensorFlow

## Conclusion

The findings from this project have potential implications for dermatologists, enabling early screening and diagnosis of skin cancers, and improving treatment effectiveness.