SoylentNews is people
SoylentNews
Digital imaging pioneer Russell Kirsch dies at 91 – TechCrunch:
Russell Kirsch, whose research going back to the '50s underlies the entire field of digital imaging, died earlier this week at the age of 91. It's hard to overstate the impact of his work, which led to the first digitally scanned photo and the creation of what we now think of as pixels.
Born to Russian and Hungarian immigrant parents in 1929, Kirsch attended NYU, Harvard and MIT, eventually landing a job at the National Bureau of Standards (later the National Institutes of Science and Technology) that he would keep for the rest of his working life.
Although he researched, coded and theorized for 50 years and even after his retirement, his most famous accomplishment is no doubt the first scanned digital image — decades before the first digital camera.
[...] This foundational work led directly to the creation of methods, algorithms and storage techniques for digital images that would inform decades of computer science. Kirsch continued his work on early AI right up until retirement, and even then continued tinkering with his idea of adaptive pixels that would enable much clearer images at lower resolutions. The idea has merit, naturally, though memory and bandwidth aren't quite the bottlenecks they once were.
Throughout his life Kirsch and his wife, who survives him with their children, were also travelers, climbers and artists. No doubt his rich life contributed to his important work and vice versa.
(Score: 2) by looorg on Sunday August 16 2020, @02:09PM (5 children)
https://www.staradvertiser.com/2020/08/13/breaking-news/computer-scientist-pixel-inventor-russell-kirsch-dies-at-91/ [staradvertiser.com]
I have been wondering about that. What would they have picked instead of squares? Why would it have been better? and why, or how, are we today suffering from square pixles? Makes perfect sense with squares if you consider images to be a grid of sorts.
(Score: 0) by Anonymous Coward on Sunday August 16 2020, @04:48PM (4 children)
Hexagons probably would be better than squares. Octo maybe better. Dodecagon probably best. The world really isn't a bunch of tiny squares all stuck together. Not at the subatomic level, atomic, molecular, or even the macro scale.
(Score: 0) by Anonymous Coward on Sunday August 16 2020, @05:12PM (2 children)
With a high enough resolution you can do that but square (or rectangular) pixels make more sense for linear scanning and rgb color arrays.
(Score: 2) by PiMuNu on Monday August 17 2020, @09:01AM (1 child)
A hexagonal grid is "line scannable" along three axes (with 120 degrees between them). I believe most line scanning routines only scan along one axis, but I am not a 3D graphics person. Hexagonal grids can be indexed using two indices just like square grids (so no difference with rgb arrays), but every other row is offset by half a grid spacing. Image is like:
https://i.stack.imgur.com/WT34G.png [imgur.com]
Not sure if it looks better however!
(Score: 2) by looorg on Monday August 17 2020, @11:15AM
That is sort of what I wondered, about or in regard to his statement, the square pixel just seems to easiest to work with. Even tho the hexagon is quite common in map-grids as far as I can recall and that the triangle is a staple of 3d-graphics. But at the time of invention at least 3d graphics, as we know it today, wasn't really a thing and hexagons require more work, building graphics with it, "pixel-by-pixel" seems a lot harder. Somehow almost not very intuitive.
(Score: 0) by Anonymous Coward on Monday August 17 2020, @08:59AM