Skin Scanner App Backend

This project is a Flask-based web application for image classification using an ONNX model. Given an input image, the application predicts whether the image is classified as "Malignant" or "Benign".

Prerequisites

To run this application locally, ensure the following are installed on your machine:

• Python 3.10
• pip

Installation

Follow these steps to set up the project locally:

1. Clone the Repository

   Clone the repository to your machine:

   git clone https://github.com/SoaresPT/Skin_Scanner_App-backend.git
   cd Skin_Scanner_App-backend

2. Create a Virtual Environment

   Set up a virtual environment to manage dependencies:

   python -m venv venv

3. Activate the Virtual Environment

   • On Windows:

     .\venv\Scripts\activate

   • On macOS and Linux:

     source venv/bin/activate

4. Install Required Packages

   With the virtual environment activated, install the necessary Python packages:

   pip install -r requirements.txt

5. ONNX Model

   ONNX models should be placed inside model. Currently defaults to: ./model/model_2.onnx.

Running the Application Locally

To start the Flask application:

python server.py

The server will be accessible at http://0.0.0.0:8000. ( or http://localhost:8000)

Testing the Live Server

The live server is accessible at: https://skinscanner.420777.xyz/predict. This will be used to send requests directly to this endpoint for predictions when using the Skin Scanner App.

Usage

To perform an inference, send a POST request to the /predict endpoint with an image file:

• Endpoint: /predict
• Method: POST
• Form Data: Image file with the key as image.

Testing with Python

You can use Python's requests module to interact with the API. Run the test_api.py. Change the url to 'https://skinscanner.420777.xyz/predict' for the live server.

Testing with cURL

curl -X POST -F "image=@path/to/your/image.jpg" https://skinscanner.420777.xyz/predict # or http://0.0.0.0:8000/predict for local

Error Handling

The application returns error messages for the following scenarios:

• No image file provided: 400 status code with an error message.
• No file uploaded: 400 status code with an error message.
• Unexpected errors: 500 status code with a detailed error message.

Docker Setup (Optional)

The repository includes a Dockerfile for containerization. You can build and run the application using Docker. However, this setup may need adjustments to suit your environment and can be ignored for initial testing. Refer to Docker documentation for building and running containers.

Skin Scanner App Repo

https://github.com/siddarthsingotam/Skin_Scanner_App