auto calendar

Inspiration

People are busy and stressed. The little things should be the last thing that they have to worry about. Science shows that keeping a calendar eases anxiety and stress, as well as helps with academic performance and managing deadlines. Especially trying to decode cryptic class schedules to find exactly what is when and where is what. A picture is said to be worth a thousand words, but now it can also be worth a thousand seconds. Keeping a calendar should be seamless, easy, encouraging to use and now it is.

What it does

We use machine learning models with deep neural networks identify, extract, and access texts from a simple picture. The text is then run through a precise and accurate machine learning model trained on over 50,000 hand labeled data points, to produce results with 99% accuracy to extract event/class information and automatically update your google calendar.

How we built it

We built it to be a website with easy to use user interfacing. The front end work is done in javascript and looks nice and pretty to give the user a streamlined platform. The back end is done in python and integrates with Google Cloud to do most of the heavy lifting for the text detection, and uses the Tensorflow library for so we can do our custom tensor computations for our machine learning model. The model is trained on a dataset of over 50,000 hand labeled points (its one of a kind and took quite a while) to produce a prediction accuracy of 99%.

Challenges we ran into

The biggest challenge was integrating the Machine Learning backend with the web front end.

Accomplishments that we're proud of

The custom dataset and model we built from scratch. Implementation of google cloud to do some of the heavy lifting. The seamless easy to use front end.