Apr 08

MIT Boosts Efficiency of Lensless Single Pixel Cameras by Factor 50

MIT Boosts Efficiency of Lensless Single Pixel Cameras by Factor 50. Original image (top), single-pixel camera image (50 and 2500 exposures, respectively; second row), and reconstructions from ultrafast sensing (third and fourth rows, 50 exposures using 100 and 20 picosecond sensing, respectively). Image: Satat et al., 2017. Today’s conventional cameras require a set of highly precise lenses and a large array of individual light sensors. This general blueprint limits the application of cameras for new uses, e.g. in ever-thinner smartphones, or in spectra outside the visible light range.
To overcome these limitations and completely rethink the basics of imaging, researchers from Rice University, Heriot Watt University and the University of Glasgow (among others) have recently developed a “compressive sensing” concept camera which uses only a single pixel and no lens whatsoever to record pictures using computational imaging. The trick here lies in the light source, which illuminates the scene using a series of defined black-and-white patterns (coded mask). Based on changes in the resulting light intensity across many exposures, the single pixel camera can then infer the position of objects and patterns in the scene.
Until recently such single-pixel systems required a large number of exposures in the range of 2000 or more. Now, Guy Satat and colleagues from MIT‘s Camera Culture lab have combined single-pixel camera with another cutting-edge technology, ultrafast femto- or picosecond light sensors. Looking not only at the intensity changes across masked illumination bursts, but also within individual bursts, the researchers are able to break up the signal into light reflected from different distances in the scene. This brings the number of required exposures down from 2500 to just 50 in the example outlined in the paper and video. Moving from a single pixel to several light-sensing pixels, placed in a defined sensor pattern, this number can be further reduced without losses in image qualilty. Continue reading

May 27

MIT: Compressive Light Field Projection System for new Glasses-Free 3D Displays

Illustration of concept. A light field projector, build using readily-available optics and electronics, emits a 4D light field onto a screen that expands the field of view so that observers on the other side of the screen can enjoy glasses-free 3D entertainment. No mechanically moving parts are used in either the projector or the screen. Additionally, the screen is completely passive, potentially allowing for the system to be scaled to significantly larger dimensions. (picture: MIT Media Lab, Camera Culture Group)Today’s glasses-free 3D displays ususally consist of dozens of devices, which makes them not only very complex, but also bulky, energy-consuming and costly. At SIGGRAPH 2014 conference, Gordon Wetzstein and Matthew Hirsch from the MIT’s Camera Culture Group presented a new approach to glasses-free 3D that is based on projectors and optical technology found in Keplerian telescopes. Their novel method for “Compressive Light Field Projection” consists of a single device without mechanically moving parts.
Because it’s relatively cheap to build with today’s optics and electronics, the presented prototype could pave the way for cinema-scale glasses-free 3D displays. Continue reading

Mar 25

Tesseract: Light Field Photography, Automated Subject Extraction and 3D Filters for your Smartphone

Tesseract Logo Remember Focii? The innovative light field capture method, developed by Kshitij Marwah at MIT Media Lab, allows users to convert ordinary cameras into light field cameras just by placing a coded mask film on top of the image sensor. Marwah has successfully ported Focii to a smartphone, and recently launched Tesseract Imaging to make the technology commercially available.
In a recent presentation at INKtalk, Kshitij Marwah presented the new product, and demonstrated three features on his modified Android smartphone (details below), live on stage: Continue reading

Aug 02

CrowdCam: Watch a Scene from any Angle, using Everybody’s Cameraphones

Events such as concerts, public performances or weddings have two things in common: Virtually everybody’s taking pictures, and it can be quite unsatisfying to be in the wrong spot. What if you could just switch your perspective to someone’s in the first row in order to get the perfect view, or even move around a scene as you like?

CrowdCam is a smartphone app concept by Aydin Arpa (MIT) and colleagues, designed to do that and more, using everybody’s smartphone cameras: The app, which is currently in development, compares photos taken at the event and estimates the different angles between camera views. It then arranges these pictures according to their relative location in the scene, giving users the ability to swipe between different points of view, while stabilizing the image and transition and keeping the image centered on the main object of interest.

CrowdCam: Watch a Scene from any Angle, using Everybody's Cameraphones (picture: Arpa et al. 2013)

In other words, the app creates a collaborative network of cameras and views, allowing you to find the best view and virtually move around in any scene.

Continue reading

Jun 26

MIT Camera Culture: Simple, Cheap Method for Light Field Photography at Full Sensor Resolution

At its current development stage, light field photography (based on microlens arrays) poses a compromise between spatial information and resolution: The more refocus or prespective a camera is required to provide, for example, the more of its sensor resolution is sacrificed. In the Lytro Light Field Camera, an 11 Megapixel sensor takes pictures that result in pictures of 1.1 Megapixels, so only about 10 % of the sensor resolution make it into the final image.

MIT Camera Culture: Simple, cheap method for Light Field Photography at Full Sensor Resolution (picture: Kshitij Marwah)MIT Camera Culture: Simple, cheap method for Light Field Photography at Full Sensor Resolution (picture: Kshitij Marwah)

As reported previously, the MIT‘s Camera Culture group has come up with a new method to capture light fields, which is both cheaper and more effective. In a new article published by MIT News, the researchers explain what their system, named “Focii”, is capable of:

At this summer’s Siggraph — the major computer graphics conference — they’ll present a paper demonstrating that Focii can produce a full, 20-megapixel multiperspective 3-D image from a single exposure of a 20-megapixel sensor.

Continue reading