Carry the Two Emerging Technologies Episode 5: Computation Imaging
Sep 25, 2025
Join experts Rebecca Willett, a computational imaging innovator, Stanley Chan, a long-distance optics specialist, and David Lindell, a high-speed imaging pioneer. They dive into how phone cameras rectify distortions and the challenge of restoring blurry images without original sensors. Discover the intricacies of adaptive optics for clear long-distance imaging and the impressive capabilities of single-photon detectors. They also explore future imaging frontiers, showcasing how new technologies can capture phenomena at staggering speeds, even light propagation!
AI Snips
Chapters
Transcript
Episode notes
Computation Makes Images Legible
- Computational imaging converts raw sensor data into human-legible images via algorithms.
- Rebecca Willett emphasizes that raw measurements require computation to form interpretable images.
Phones Rely On Hidden Computation
- Modern phone cameras use computational imaging to correct lens design and low-light issues.
- Rebecca Willett notes computation undoes distortions and drastically improves night photos.
Regularization Resolves Ambiguity
- Classical computational imaging uses models of noise and geometry to search for images that fit data.
- Regularization encodes prior beliefs like smoothness to select plausible reconstructions.