API.md

API Documentation

Overview

The CNN Image Classifier provides a RESTful API built with FastAPI for image classification. The API allows users to upload images and receive predictions with confidence scores.

Base URL

http://localhost:8000

Endpoints

1. Welcome Endpoint

GET /

Returns a welcome message and basic information about the API.

Response

{
    "message": "Welcome to the CNN Image Classification API",
    "version": "1.0.0",
    "endpoints": {
        "predict": "/predict"
    }
}

2. Prediction Endpoint

POST /predict

Classifies an uploaded image and returns the predicted class with confidence scores.

Request

• Method: POST
• Content-Type: multipart/form-data
• Body:
  • file: Image file (supported formats: JPG, PNG)

Response

{
    "class": "airplane",
    "confidence": 0.95,
    "top_3_predictions": [
        {
            "class": "airplane",
            "confidence": 0.95
        },
        {
            "class": "bird",
            "confidence": 0.03
        },
        {
            "class": "cat",
            "confidence": 0.02
        }
    ]
}

Error Responses

1. No File Uploaded

{
    "error": "No file uploaded"
}

2. Invalid File Type

{
    "error": "Invalid file type. Supported formats: JPG, PNG"
}

3. Processing Error

{
    "error": "Error processing image"
}

Example Usage

Using cURL

curl -X POST "http://localhost:8000/predict" \
     -H "accept: application/json" \
     -H "Content-Type: multipart/form-data" \
     -F "file=@path/to/your/image.jpg"

Using Python

import requests

url = "http://localhost:8000/predict"
files = {"file": open("path/to/your/image.jpg", "rb")}
response = requests.post(url, files=files)
print(response.json())

Using JavaScript

const formData = new FormData();
formData.append('file', fileInput.files[0]);

fetch('http://localhost:8000/predict', {
    method: 'POST',
    body: formData
})
.then(response => response.json())
.then(data => console.log(data))
.catch(error => console.error('Error:', error));

Rate Limiting

Currently, there are no rate limits implemented. However, please be mindful of server resources when making requests.

CORS

The API supports CORS and can be accessed from any origin. For production deployment, you may want to restrict this to specific domains.

Error Handling

The API uses standard HTTP status codes:

• 200: Success
• 400: Bad Request
• 415: Unsupported Media Type
• 500: Internal Server Error

Security

• File size limit: 10MB
• Supported file types: JPG, PNG
• Input validation for all requests
• Error messages are sanitized to prevent information leakage

Future Enhancements

1. Authentication
2. Rate limiting
3. Batch processing
4. Additional model endpoints
5. Model versioning