Skip to content

Latest commit

 

History

History
160 lines (107 loc) · 4.15 KB

README.md

File metadata and controls

160 lines (107 loc) · 4.15 KB

Image Processing and Organization Tool

Overview

This project is an advanced image processing and organization tool that utilizes machine learning models for various image-related tasks. It features a backend powered by Hugging Face's transformers.js, a frontend for user interaction, and a Bash script to automate image renaming based on generated captions.

Features

  • Image Captioning: Automatically generates captions for images.
  • Object Detection: Identifies objects present in an image.
  • Emotion Detection: Recognizes emotions from facial expressions.
  • OCR (Optical Character Recognition): Extracts text from handwritten or printed images.
  • Image Upscaling: Enhances image resolution.
  • Automated Image Renaming: A Bash script renames images based on their generated captions.

Backend

The backend is built using Node.js and Express, utilizing the @huggingface/transformers library for AI-powered image processing.

Key Components

  • Express Server: Handles API requests.
  • Multer: Handles file uploads.
  • Hugging Face Models: Loads and runs various image processing models.
  • Dynamic Route Generation: Each model has its own API endpoint.

Available API Endpoints

Endpoint Functionality
/generate-caption Generates captions for images
/detect-obj Detects objects in an image
/detect-emo Recognizes facial emotions
/generate-ocr Extracts text from images
/upscale Enhances image resolution
/routes Lists all available endpoints

How to Run the Backend

  1. Install dependencies:

    npm install

  2. Start the server:

    node app.js

  3. The server runs at http://localhost:5100

Image Renaming Script

A Bash script automates renaming images based on AI-generated captions.

How It Works

  • Sends an image to the API.
  • Retrieves the generated caption.
  • Renames the image using the caption.
  • Supports parallel processing for faster execution.

Usage

./rename_images.sh -p 5 -sr _ -r responses.txt -sf image.jpg

Options

Flag Description
-p Set concurrency level (default: 3)
-sr Space replacement in captions
-sf Process a single file instead of a directory
-r Save API responses to a file
-d Set search depth for images in directories

Frontend

The frontend is a React.js application that provides an intuitive UI for users to upload images and receive processed outputs.

Features

  • Upload images for processing.
  • Select the desired processing task.
  • View results with a clean UI.
  • Monitor the processing status.

Screenshot Preview

Action Selection Caption Generation Emotion Detection Object Detection OCR Upload

How to Run the Frontend

  1. Navigate to the frontend directory:

    cd frontend

  2. Install dependencies:

    npm install

  3. Start the development server:

    npm start

  4. Open http://localhost:3000 in your browser.

Installation

Prerequisites

  • Node.js & npm
  • Bash (for script execution)

Steps

  1. Clone the repository:

    git clone https://github.com/Sahil-958/ims.git

  2. Set up and run the backend.

  3. Run the frontend.

  4. Use the script to rename images if needed.

Future Enhancements

  • Implementing user authentication.
  • Extending support for more image transformations.
  • Adding a database to store processing history.