Using Python’s Numpy is easy to draw how an average Chinese character would look like. I took the 2000 most common characters hanzidb.org, drew them onto images using PIL and calculated it’s average with Numpy. While I can’t read a single word of Chinese, this is what struck me the most:
• There are four main lines: top, bottom, middle, and left.
• The bottom line increases in intensity towards the right, likely a remanence of Chinese being a brush script.
• The darkest area occurs at the middle bottom.
• We can appreciate a few short strokes on the left top side.
If you have any other insights or ideas send us a comment!

This site uses Akismet to reduce spam. Learn how your comment data is processed.