-
-
Presentation Screen
-
Home Screen
Inspiration
As University students we often get assigned many readings, and due to time constraints, it results in us rushing our readings, and not learning much. We thought we would be able to improve our learning experience, if we could summarize our readings, and have them converted into a lecture/presentation format. While thinking about this idea, we started wondering how to enhance the learning experience.
What it does
The web app allows you to paste an article/reading, or take a picture, and then summarizes that piece of text using Natural Language Processing. It then converts that summary into a presentation format, with slides and bullet points. On top of that, a narration of the generated slides can be played, the narration is in Kevin Hart's voice, which was generated using Microsoft Azure. In the main portal, the user is also able to take notes, and look up any words in a dictionary via oxford dictionary api.
How I built it
Backend was built using flask, angular was used for the front-end. Python's NLTK library was used for text summarization of the text, google cloud vision API was used for the image-to-text generation, and azure was used for the text-to-speech.
Challenges I ran into
Creating the dataset for the voice model, summarizing text, designing UI
Accomplishments that I'm proud of
Successfully generating Kevin Hart's voice
What I learned
Using Azure text-to-speech Using GCP Image-To-Text Playing audio in browser, sending audio from server to client Kevin Hart's voice can get annoying after transcribing 100's of audio files
What's next for Smartr
As of right now we only have Kevin Hart's voice generated, however with more time we can add other voices. Also, we want to make our program smarter, so we would like to add features such as question generation based on slides etc. With tweaks and improvements, we see this project as something we could launch to many students in different schools soon.
Log in or sign up for Devpost to join the conversation.