On Apple's heels this morning, a $ 600 million deal was signed to acquire IP, talent and licenses from Dialog Semiconductor in Europe. It has also confirmed another acquisition of a smaller start-up in the region.
Apple has purchased Spektral, a data-based company based in Denmark that worked with segmentation technology, a more efficient way to "cut out" numbers from the background in digital images, allegedly for around $ 30 million.
This type of technology can, for example, in magnified reality, but also for more common applications such as school photographs (which really were the first market, started targeted, by 2015).
Rumors about the deal began to surrender yesterday, first in Danish newspaper Børsen. We traveled to Apple, and the company has now confirmed the agreement with its standard statement: "Apple is buying small technology companies from time to time, and we do not usually discuss our intent or plans."
From what we understand, the acquisition took place a while ̵
Others related to the company – including the other founder, Henrik Paltoft – have not updated their profiles, so it is unclear how many others have joined. The stock exchange reports that the agreement includes the company's engineers and is in the region of NOK 200 million, which is equivalent to approximately $ 31 million.
Spektral started life as CloudCutout, based on algorithms from Jansens PhD. Start-up initially set its product as a cheaper and more efficient "green screen" technology to remove primary images from their usual (usually green) or standard pattern backgrounds, with early iteration of the product built by training the system over 100,000 professional clippings.
Spectral's first application may have been the pretty retro world of school pictures, but what's most remarkable here is what Spektral can contribute to Apple's image processing. It's not just applications that Apple has yet released but to improve the quality of those already on the market, from older products like PhotoBooth to ARKit, the company's platform for mobile development.
Segmentation can help add live filters to human characters in an image, but can also be effective in occlusive AR environments behind figures to make digital AR content interact with people's position.
Spectral segmentation technology can also run on mobile phones, enabling a faster and more efficient way to process AR images directly on devices.
"To provide high quality cuts, the core of our engine utilizes the latest advances in spectral graft theory and neural networks. The calculation of pixel transparencies (alpha channel) for a single image involves solving multiple large-scale equation systems, as well as carrying Multiple feed passes in our neural networks, "the founder said when launching the first time.
"We set the problem of determining an alpha channel in a picture as a machine tutorial. Compared with regular chroma tasting, this allows us to evaluate a much wider range of backgrounds the model will learn, ie texture representations from existing training data. "
Data vision has been a central focus for Apple for a while now. The iPhone and Mac giant have made more than 40 acquisitions in Europe over the last 10 years – I guess we can still have a little hunting – and a number of acquisitions in Europe and elsewhere have been in the field of data. They have included Akonia Holografics, InVisage Technologies, Regaind, Vrvana, SensoMotoric Instruments, Indoor.io, Flyby Media, Emotient, Faceshift, Metaio, Polar Rose and more.
Additional Report Natasha Lomas.