Tesseract promises “a new camera technology that allows you to perform SLR-like depth of field effects, digital refocusing, 3D color filters and much more on your mobile phone.” Noting its ability to generate separate layers, TechCrunch says, “It’s part RAW, part PSD but straight from your mobile device’s camera.”
Post-capture refocusing still strikes me as a parlor trick, but image segmentation to enable better extraction, compositing, and depth of field effects seems highly useful. You can get a bit more info from founder Kshitij Marwah’s recent talk.
Meanwhile, the company has released FOCII, a printable transparency mask that lets you “take any existing DSLR and convert into a light field camera for post capture refocusing with a simple $1 filter!”
hi (again) John,
If you want to research and/or pursue image segmentation, then I recommend that you contact Prof. Andrew Bangham (https://www.uea.ac.uk/computing/people/profile/a-bangham). As commented in part the other day, Andrew was a founder of two companies (Segmentis and Fo2PiX) in the U.K., which looked at sub-dividing digital images in a way which allowed for compositional segmentations. For facial recognition and digital artwork applications (amongst other things).
I visited with Andrew and his staff in around 2005, at what was then the HQ of Fo2PiX near Cambridge. Later it seems he took his academic career more seriously than commercial activities. But he was a true pioneer on these topics. And can tell you that some of the algorithms developed were not – at that point in time – practically computable. But, they went beyond what is demonstrated in the TED video you link to. (I still have some documentation on the original work which made up the intellectual property of Segmentis Ltd. – but I’d have to search a bit to find it …)
… part of that would read more informatively as: “… that some of the algorithms developed were not – at that point in time – practically computable with a professional-level PC … “
I have been just shocked and thrilled with all this super high tech stuff that is being done with cell phones and all kinds of small jpg’s, and they are all small jpg in order for these processes to work, and all with relatively low resolution. So what? All that is fun and cute but has no practical use whatsoever, and is not preliminary research to a higher plane. Are cell phone sales that low that all this extra bs is being added? REALLY? If the pieces will combine for facial recognition or something like that, ok, but otherwise it is a waste of effort. Instead of buying from me, people get temporary gratification from this type of thing. Barry Blanchard says that I can do a high res print from the quadcopter, OK. What can I do with this Teseract? and the other stuff like it.