Caption IT: Beautiful captions Suggestions by AI:sparkles:

Sucked 😒 by searching some nice captions for your post online. Bored of posting random captions with your images.

Use caption IT to suggest some captions from the internet that are suitable to your image through AI.

You can also get your captions based on your given keywords.

How It Works

The algorithm uses machine learning and natural language processing to process. It all works in three major steps:

• It extract all the objects from the image (eg. flower, sun).
• It uses natural language processing to find synonyms of all the objects detected in the image.
• Then it use an external api to fetch quotes related to all the words.

Technology Used

• FrontEnd : React, ReactRouter, Sass
• Backend API: Flask
• Object Detection: TinyYoloV4 with ImageAI
• NLP: NLTK
• Quotes Api: Quotable.io

Run Locally

Clone the project

  git clone https://github.com/piyushi2227/Caption-IT

Go to the project directory

  cd rock-paper-scissors-online

Go to backend and install dependencies

cd Backend
pip install -r requirements.txt

Start the server

python server.py

Install frontend dependencies

  cd ..
  cd Frontend
  npm install

Start the Frontend server

npm run start

Future Scope

• The application is slow as it does many high performance things at the same time and due to that it does effect the user experience.
• The object detection might be replaced or kept with image captioning to extract keywords from image.

Author

• @piyushi2227

Feedback

For feedback, email patidarpiyushi@gmail.com.