Computer vision’s potential expansion across all of science and engineering
The author points to robotics and medical imaging as examples of how domains with visual data could become computer vision subfields, each with its own challenges and solutions.
TLDR
A post argues that computer vision needs to expand across science and engineering to stay relevant. The author warns that a narrower focus on image, video and 3D understanding and reconstruction could become “mostly done” and shrink in relevance. They describe seeing people from physics, chemistry and physiology at CVPR as a hopeful sign, and wish for it to become an all-science-and-engineering festival.
Combined views
30
1 Source, first seen ago