DermAI Skin Lesion Classification

Project Overview

DermAI is a Flask-based application designed to classify skin lesions using ML models that supports two model options which are CNN (image-only) and CNN + Metadata (image with metadata)

---

Important Project Directory

├── app.py                           # Backend Logic/Model Logic
├── templates/
│   └── index.html                   # Web UI
├── data/
│   ├── test/                        # Directory for test images
│   └── test.csv                     # Metadata for test images
├── models/
│   ├── best_cnn_model_traced.pt            # CNN model
│   └── best_cnn_metadata_model_traced.pt   # CNN model with metadata
├── static/
│   ├── group.png                   
│   └── styles.css                   
├── Dockerfile                      
├── cloudbuild.yaml                  # Google Cloud Build configuration
├── requirements.txt                 # Python dependencies
└── README.md                    

---

Requirements

• Python 3.8+
• Flask 2.x
• PyTorch 2.0+
• Docker (optional)
• Google Cloud Build and Cloud Run (for deployment)
• Dependencies listed in requirements.txt

---

Installation and Usage

Step 1: Clone the Repository

git clone https://github.com/sofiahanim/skinlesionclassificationDermAI.git
cd skinlesionclassificationDermAI

Step 2: Install Dependencies

pip install -r requirements.txt

Step 3: Run the Application

python app.py

The application will be accessible at http://127.0.0.1:5000

---

Features

• Model Options:

  • CNN provide basic classification using images only
  • CNN + Metadata enhanced classification with images and additional metadata such as age, sex, anatomical site

• Predictions:

  • Upload skin lesion images (PNG/JPG/JPEG)
  • Choose the model type and view predictions with confidence scores

• Test Data Browser:

  • Search and view test data in a table format
  • Preview and download test images

---

Deployment to Google Cloud Run via Cloud Build

Step 1: Configure Cloud Build in Google Cloud Console

1. Navigate to Cloud Build in the Google Cloud Console
2. Create a new trigger:
   • Connect GitHub repo
   • Set the trigger to build on branch main
   • Use cloudbuild.yaml file for build configuration

Step 2: Deploy to Cloud Run

1. Once the build is complete, go to Cloud Run in the Google Cloud Console
2. Deploy container image created by Cloud Build

---

cloudbuild.yaml

steps:
  - name: 'gcr.io/cloud-builders/docker'
    args: ['build', '-t', 'gcr.io/$PROJECT_ID/skin-lesion-classification', '.']

  - name: 'gcr.io/cloud-builders/docker'
    args: ['push', 'gcr.io/$PROJECT_ID/skin-lesion-classification']

  - name: 'gcr.io/cloud-builders/gcloud'
    args:
      - 'run'
      - 'deploy'
      - 'skin-lesion-classification'
      - '--image'
      - 'gcr.io/$PROJECT_ID/skin-lesion-classification'
      - '--platform'
      - 'managed'
      - '--region'
      - 'us-central1'
      - '--allow-unauthenticated'

---

Notes

• Update the project-id and region placeholders in cloudbuild.yaml with actual values
• The application will be accessible at URL provided eg: https://skin-lesion-classification-[id].run.app

---