Results 1 to 3 of 3
Thread: Cartooning
- 01-10-2011, 01:14 PM #1
Member
- Join Date
- Dec 2010
- Posts
- 20
- Rep Power
- 0
- 01-12-2011, 09:48 PM #2
Yeah, its most likely using a series of filters and algorithms:
The image is probably converted to grayscale, and then contrast is adjusted until mid tones are washed out. Then an edge detection algorithm traces the edges of objects. Some algorithm can probably pick out the basic shape of a head/face/body from the resulting outline. Once this is achieved, the original image can have its central figure extracted. Posterizing might be used in here somewhere too. Another sweep with contrast adjustments, and facial component locators finds the basic shapes and locations of mouth, eyes, hair, etc...
From here, the located objects are probably matched to a predefined database of shapes (hairstyles, eye shape, mouth shape, etc...). The face is redrawn using these pre made shapes, colors either sampled from averages in the original image or from user input. The completed composite image is returned to the user!
Sounds like a fun weekend project :D
- 01-13-2011, 09:16 AM #3
Member
- Join Date
- Dec 2010
- Posts
- 20
- Rep Power
- 0


LinkBack URL
About LinkBacks
Reply With Quote
Bookmarks