Hi! [Christopher Phillips] > For each 4x8 tile in the source image, compute some measure > of the nature of the colours in that tile; for example > the means of R,G&B and the 3x3 covariance matrix. > You then have an 80x50x12 'image' of what sorts of colours > reside in each tile. > [...] Thanks. Sounds like it will work. Will try it out today. > This is of course ignoring the background colour for now; > perhaps sanest to pick one in advance of generating > the smaller palettes, and allocate it to the sub-palette > that would least miss control of it's 16th colour, > and treat all the other sub palettes as already having > a 17th entry that is the background colour. Yes, I also pre-selected the background color. In the current prog, I did it by doing the whole color selecting process without restrictions on the number of colors per blocks, and collected colors from each blocks where the number of colors (using some reduction based on the spread of the included colors) was higher than 3. After the process, the background was selected by the color that was found most frequently in the above blocks (ie. in the highest number of blocks). With this, the algorithm 'enhances' color resolution with the background color, just in the blocks where it is needed. After this process, the algorithm starts selecting color triads for all blocks, classifying colors by excluding those that are below certain distance from the pre-selected background color. Should be neccessary to do something like this; will have to think it over again. > Hmm. Thinking a little more, a better alternative might > be to quantise the 80x50 image down to 16*(16/3) = 80 > sets; you can then generate 80 independent three-entry > palettes, and group them into 16 sets of 5 palettes > each for packing into the sub-palettes. Will think over both again. Thanks :-). >[...] > I was under the impression that the main objection people had > to discussions relating to Jeri's project was on the grounds > of the majority of them being wish-lists and dreaming, instead > of technical discussions about implementation. Because, you're right. (MagerValp has told that the best). I'm playing with the thought that past week's thread is a typical example of confusing reason with result. Best regards, L. - This message was sent through the cbm-hackers mailing list. To unsubscribe: echo unsubscribe | mail cbm-hackers-request@dot.tml.hut.fi.
Archive generated by hypermail 2.1.1.