Dear Professors/Friends,
Since research gate website has a capability of putting discussion, I send a question.
In color space models naming RGB, L*a*b*, HSV, HSI, XYZ, CMYK, YIQ, YUV, YCbCr, and L*u*v*, when we put the color digital image and cluster its data, the color quantization or reduction is done, but which one hard partitioning or fuzzy partitioning? FCM based algorithms or K-Means based algorithms are better? I mean do you think these named color spaces will provide overlapped color date or not?
Best Regards,
Farshid Keivanian
MSc in Electronics Engineering
University of Birjand
Iran
Mr Keivanan, I have used L*a*b* as feature space to perform color identification but I have never clustered it. But on observing those feature spaces, probably I will go for Fuzzy C means clustering because, I have observed smooth transition of colors across a* and b* components in that color space.
- Badges
- Science topic
- Similar topics
- Force