TERI, almost IRL Blade Runner movie image enhancement tool
Researchers at the University of South Florida introduce TERI, an image processing algorithm inspired by Blade Runner, reconstructing hidden objects in photos using shadows. Potential applications in self-driving vehicles and robotics.
Read original article
Researchers from the University of South Florida have developed an image processing algorithm that can reveal hidden objects in a 2D photograph, similar to a concept from the movie Blade Runner. The algorithm can reconstruct obscured objects and rooms outside the frame by analyzing shadows cast into the shot by objects not visible in the original photo. This technology, named TERI, could have applications in various fields such as self-driving vehicles and robotics, although it may take a decade or two to be robust enough for widespread use. The team's work has been published in the journal Nature Communications, and MATLAB scripts for reproduction are available on GitHub. The algorithm aims to provide a more complete 3D picture of scenarios by utilizing shadows to infer the shapes of objects and rooms not initially visible in the image.
Related
HybridNeRF: Efficient Neural Rendering
HybridNeRF combines surface and volumetric representations for efficient neural rendering, achieving 15-30% error rate improvement over baselines. It enables real-time framerates of 36 FPS at 2K×2K resolutions, outperforming VR-NeRF in quality and speed on various datasets.
New device uses 2D material to up-convert infrared light to visible light
Indian Institute of Science researchers create a device converting infrared to visible light using gallium selenide layers. Efficient up-conversion demonstrated, promising compact and improved infrared imaging applications without bulky sensors.
Moving objects precisely with sound
EPFL researchers use soundwaves for precise object manipulation, advancing drug delivery. Wave momentum shaping navigates objects in dynamic environments, offering noninvasive biomedical applications and potential for micro-level cell manipulation. Nature Physics publication showcases groundbreaking research.
Ghosts in the (Macintosh) ROM (2012)
A group at NYC Resistor found hidden images in Apple Mac SE ROM from 1986. They decoded distorted data, revealing pictures and assembly code, extracting four images and hinting at audio data.
Subwavelength imaging using a solid-immersion diffractive optical processor
Researchers developed a solid-immersion diffractive optical processor for subwavelength imaging using deep learning. The system magnifies images, reveals subwavelength features, and operates across electromagnetic spectrum for various applications.
0 commentsRelated
HybridNeRF: Efficient Neural Rendering
HybridNeRF combines surface and volumetric representations for efficient neural rendering, achieving 15-30% error rate improvement over baselines. It enables real-time framerates of 36 FPS at 2K×2K resolutions, outperforming VR-NeRF in quality and speed on various datasets.
New device uses 2D material to up-convert infrared light to visible light
Indian Institute of Science researchers create a device converting infrared to visible light using gallium selenide layers. Efficient up-conversion demonstrated, promising compact and improved infrared imaging applications without bulky sensors.
Moving objects precisely with sound
EPFL researchers use soundwaves for precise object manipulation, advancing drug delivery. Wave momentum shaping navigates objects in dynamic environments, offering noninvasive biomedical applications and potential for micro-level cell manipulation. Nature Physics publication showcases groundbreaking research.
Ghosts in the (Macintosh) ROM (2012)
A group at NYC Resistor found hidden images in Apple Mac SE ROM from 1986. They decoded distorted data, revealing pictures and assembly code, extracting four images and hinting at audio data.
Subwavelength imaging using a solid-immersion diffractive optical processor
Researchers developed a solid-immersion diffractive optical processor for subwavelength imaging using deep learning. The system magnifies images, reveals subwavelength features, and operates across electromagnetic spectrum for various applications.