Suppose I have an image and want to show the distribution of colours in that image. Obviously, I can sort the pixels in a line and be done. But what if I wanted to preserve the original photo's dimensions, while minimising the colour distance across neighbouring pixels in both dimensions?
[#]graphics #computationalimaging #art
=> More informations about this toot | More toots from addison@nothing-ever.works
@addison something something cosine transform?
=> More informations about this toot | More toots from buherator@infosec.place
@buherator@infosec.place How would that be applied? My goal is to simply rearrange the pixels, this seems more like signal processing.
=> More informations about this toot | More toots from addison@nothing-ever.works
@addison TBH I just vaguely remember that cosine transform is useful for compression because it tends to group similar values together, and it also works in 2D so it was the first thing that came to my mind, but it's been a looong time since I touched anything related.
=> More informations about this toot | More toots from buherator@infosec.place This content has been proxied by September (3851b).Proxy Information
text/gemini