Smart Storyteller
An NLP based storytelling from images
Overview
This project aims to create a deep learning model for the purpose of automatically generating captions for images and thereafter performing storytelling with the help of that caption. The algorithm involved will generate a suitable caption for the image as a one-liner and generate a specific story based on the keywords from the caption as input. Being able to automatically describe the contents of an image after image processing, using properly formed English sentences is rather a very challenging task. Although, it proves to have various viable and effective applications in our day-to-day lives. For instance, visually impaired people can be made to understand what the image is about. Creating news from an image is another application using this principle. Creating stories, poems or plays in the style of someone who’s either no longer alive or wishes to create his/her works in a more timely manner. It can even be used for the purpose of script writing. Hence, this project encompasses the creative method of storytelling by studying an image by detecting and understanding its contents using deep learning and generating human-like stories thus proving it to be a viable approach for various applications.
Technology used
Python, Pandas, Numpy, InceptionV3, GPT2