kshitizb/pixelscan repository overview

⁠Pixelscan - A FastAPI-Based Image Classification API

⁠Overview

pixelscan is a FastAPI application designed for image classification using multiple deep learning models (DCNN.h5, mobilenet_model.h5, and xception_model.h5). This API allows users to upload images and receive a prediction on whether the image is "Real" or "Fake."

This containerized application is built with Python 3.12.3 and runs efficiently on Docker. It can be deployed on Raspberry Pi 5, cloud servers, or local machines.

⁠Features

✅ FastAPI-based web application
✅ Supports image upload and classification
✅ Uses three deep learning models for ensemble prediction
✅ Lightweight Dockerized solution
✅ Prebuilt Docker image available on Docker Hub for easy deployment

⁠Run Locally with Docker

⁠1. Pull the Docker Image

To download the latest version of pixelscan from Docker Hub, run:

docker pull kshitizb/pixelscan:latest

⁠2. Run the Container

Start the container and expose it on port 8000:

docker run -d -p 8000:8000 --name pixelscan --restart unless-stopped kshitizb/pixelscan:latest

⁠3. Access the API

Once the container is running, open a browser and go to:

http://<your-server-ip>:8000

Or test via FastAPI interactive docs:

http://<your-server-ip>:8000/docs

⁠Endpoints