Colour conversion matrices - Tutorial by Andy Rowlands

alanr0

Veteran Member
Messages
2,611
Solutions
15
Reaction score
1,784
Location
Harlow, UK
Last edited:
Thank you for posting this.

I did not read the whole paper, just a remark on what I see on pp. 8-9 there. A variation of the standardized method B of ISO 17321-1 there involves (1) computing a color matrix T by eq. (14) which is a standard way to find an approximate solution of a system which may not be even solvable (least squares) and (2) optimize T by minimizing ΔE using the nonlinear optimization technique recommended by ISO 17321-1. The latter publication is behind a paywall and my university does not seem to be subscribed.

One may wonder if the optimization above is done with a good choice of an L^2 product. On the other hand, (2) probably takes care of this.
 
Thank you for posting this.

I did not read the whole paper, just a remark on what I see on pp. 8-9 there. A variation of the standardized method B of ISO 17321-1 there involves (1) computing a color matrix T by eq. (14) which is a standard way to find an approximate solution of a system which may not be even solvable (least squares) and (2) optimize T by minimizing ΔE using the nonlinear optimization technique recommended by ISO 17321-1. The latter publication is behind a paywall and my university does not seem to be subscribed.

One may wonder if the optimization above is done with a good choice of an L^2 product. On the other hand, (2) probably takes care of this.
I don't have any more details of the ISO 17321-1standard.

FWIW, Keith Borrino (GoPro, with co-authors from Rochester Institute of Technology) has published a sensitivity analysis comparing the results of the test chart based method and the spectral characterisation method from ISO 17321-1.

 

Keyboard shortcuts

Back
Top