Bart Wronski 🇺🇦🇵🇸(14 pages, Siggraph Asia, ffs)
nicebyte@BartWronski this describes like... the majority of stuff that gets published? :)
@nicebyte heh, depends :)
I had a low opinion of computer vision conferences (incremental, focused on benchmarks) and, conversely, "most Siggraph papers are truly innovative, even if not useful", but recently there has been some computer vision style creep into Siggraph papers as well...
I had a low opinion of computer vision conferences (incremental, focused on benchmarks) and, conversely, "most Siggraph papers are truly innovative, even if not useful", but recently there has been some computer vision style creep into Siggraph papers as well...
Sebastian Sylvan@BartWronski @nicebyte siggraph has high concentration of cool ideas. But 95% if it will never see the light of day in any real application.
@ssylvan @nicebyte I'm ok with 100% impractical, as long as it's inspiring and novel. This is the goal of publishing, research, and academia.
I am not ok with incremental (non)improvements beating some benchmarks without much real innovation. Those won't see the light of day either and won't inspire anyone.