HackCursive | Devpost

Testing out our whiteboard!

Inspiration

It seems that students today are being taught cursive less and less. We wanted to create a solution to help people learn a dying yet ever so common style of handwriting in a unique way.

What it does

HackCursive helps anyone with access to a webcam to learn cursive. The application connects to the user's webcam and outputs a mirrored view of their camera. On top of this view is an overlay of cursive text that the user can trace to practice their cursive. To begin, the program works by opening up a screen that activates the "whiteboard." This whiteboard is created with our front facing camera. At the top of the screen are different colour options, so if you wish the change the colour of your marker you can simply hover over the colour you want. The colour that you calibrated it to detect will have a circle around it. This circle will show you the path you are drawing.

How we built it

We built HackCursive using OpenCV and Python. We implemented it with the Google Vision API to detect if the letters drawn out by the user were correct.

Challenges we ran into

Some challenges we ran into were calibrating the colours once the user opened the application. We were able to solve this by placing a box on the screen. There were two smaller paths to this process - reading the RGB value of the pixels in the image, and the other was getting the image from the webcam. Getting the value of the pixels took a long time since we had to import PIL, which is a python library that allows you to manipulate images. Then, we cropped the image down to an 100x100 square in the middle. Then, we used an array to store the RGB value with a for loop. Then, we sorted the values. To find the range of values we needed for our object, we went from the first value in the array to the last value in the array.