A rather secretive startup from Hollywood, Florida, recently made headlines for raising a spectacular investment for their vision of the next generation of Virtual Reality. Big names like Google, Qualcomm Ventures, Andreessen Horowitz, and others have put together the sum of 540 million US-Dollars for a company called Magic Leap, but the public isn’t even sure what the company is working on.
The official press release reads: “Magic Leap is going beyond the current perception of mobile computing, augmented reality, and virtual reality. We are transcending all three, and will revolutionize the way people communicate, purchase, learn, share and play.”
…and Magic Leap’s website doesn’t provide many details either.
The company is reportedly working on Dynamic Digitized Lightfield Signals” (Digital Lightfield, in short), a “biomimetic” technology that “respects how we function naturally as humans”. What that means precisely, the company doesn’t explain. However, Technology Review has dug up some interesting patent applications by Magic Leap which may give us a glimpse into what convinced their investors: Continue reading →
Calibration is an important part of light field photography: Image processing and image quality can be significantly improved when the physical properties of the camera are known. More specifically, geometric information about the microlenses in a microlens-array-based light field camera can help create more precise depth maps with fewer errors.
Yunsu Bok and colleagues from the Korean Advanced Institute of Science and Technology (KAIST) have devised a new method for geometric calibration which – in contrast to conventional methods – does not rely on processing sub-aperture images. Instead, they extract line features and compute a light field camera’s geometric parameters directly from RAW images. Continue reading →
Jan Kučera has recently released a suite of software updates for his Lytro Meltdown tools, the Lytro Compatible Viewer (updated to version 126.96.36.199), the Lytro Compatible Communicator (new version: 188.8.131.52), and the Lytro Compatible Library (new version: 184.108.40.206).
Updates include a 3D mesh view from depth maps for the Viewer, improved demosaicing, and user manuals. The library has received accessors for well-known components in light field packages, dedicated classes and methods for easier access to sub-aperture and individual microlens images.
Earlier this year at Augmented World Expo, Nvidia researcher Douglas Lanman gave a talk about Near-Eye Light Field displays, i.e. electronic glasses which allow users to experience both 3D and depth. When asked about Augmented Reality (AR) applications during the discussion, Lanman noted that creating a set of transparent glasses that would also include microlenses (or something equivalent) but still allow “normal” see-through vision, was a real challenge. He very briefly teased “pinlight displays”, which were to be presented at the same conference, but no further information could be found online.
In the Emerging Technologies section of the Siggraph 2014 conference (10-14 August 2014), Adam Maimone and colleagues from the University of North Carolina at Chapel Hill and Nvidia will be presenting their new invention in a talk entitled “Pinlight Displays: Wide-Field-of-View Augmented-Reality Eyeglasses Using Defocused Point-Light Sources”. Continue reading →
With today’s light field sensors, extracting 3D stereo images from light field recordings typically results in a lowered effective image resolution – but that limitation may soon be history: Sony has developed a novel sensor design with overlapping pixels in two layers, that will allow 3D output without the typical decrease in image resolution. In Sony’s recently granted US Patent, Nr. US20140071244, author Isao Hirota introduces a dual level microlens array setup in combination with a sensor that consists of two layers of light sensitive pixel grids – front-facing and back-facing grids that are rotated at, for example, 45 degrees.
The described configuration allows different neighbouring pixels to share the same information from a single microlens while being allocated to either the left or right stereo views, resulting in higher-resolution 3D stereo output from a single-lens, single-sensor device (i.e. a “monocular 3D stereo camera”).