Scribble Score

• 

  Line of Best Fit for a line of relatively neat handwriting

• 

  Line of Best Fit for a line of relatively messier handwriting

• 

  sample test of detecting words and spaces

I've always had messy handwriting growing up, which resulted in consequences such as lower grades and illegible class notes. I never found a way to properly practice improving my penmanship, so I concluded that a handwriting improvement app would be useful for children who have the same problem that I did at their age.

Our project scans your handwriting and processes it to both read the text, and rate the legibility of the handwriting based on parameters such as consistent spacing, and how well it fits on a straight line.

We utilized the Sobel filter to detect edges and important details in the image that was scanned in, then found a line of best fit to compare the pixels to determine how straight the handwriting was. We created bounding boxes around words, calculating the variation in distance between clusters to determine how consistent the spacing in the handwriting was.

Some challenges we ran into were debugging the backend AI code to scan and analyze the handwriting properly and figuring out how to connect the backend Python models to the frontend React. Since none of us had much experience with front-end, we had to learn on the spot and consult mentors for help.

For many of us, this was the first hackathon we have participated in and been able to complete a project, so the experience of having completed a project such as this in such a short timeframe is very rewarding.

Built With

• css
• flask
• github
• javascript
• matplotlib
• numpy
• opencv
• python
• react
• tesseract