How our smartphone photos could be given a Spectral Edge
PUBLISHED: 23:13 25 July 2018 | UPDATED: 23:31 25 July 2018
Iliffe Media Ltd
The Bradfield Centre company’s pioneering image fusion processing can improve mobile phone pictures, TV displays, security camera footage and more
With more than 1.5 billion smartphones sold each year, and millennials expected to take an incredible 25,000 selfies during their lifetime, imaging is big business.
Spectral Edge, a spin-out from the Colour Lab at the University of East Anglia, is gathering momentum in the market with its image fusion processing technology that can significantly improve photographs and videos.
It is one of the success stories at The Bradfield Centre, the deep tech collaborative space on Cambridge Science Park that has celebrated its first anniversary.
Rhodri Thomas, CEO of Spectral Edge, told the Cambridge Independent: “The IP was created by Professor Graham Finlayson at UEA. It uses very clever algorithms and mathematical solutions to fuse two different versions of an image together, very rapidly – using real-time processing – with no artefacts and nasties left behind, which is what you normally get when you merge two images together.”
The pixel-level, embeddable technology can be implemented in software or in silicon, giving it wide application. In addition to the smartphone market, Spectral Edge’s technology could be used in security and surveillance cameras, on webcams, on TVs and monitors and in the automotive industry for rear-view cameras or dashcams.
An oversubscribed second funding round of £1.5million in March 2016 led by Parkwalk Advisors and IQ Capital was followed in April this year with a further £4million from them.
Rhod, who joined in February 2017, said: “There were four or five people two years ago. Now we have 16. We have a dozen in the R&D team to productise this technology. We have a professor, four PhDs in image processing and five MScs in image processing. We’ve built a really high-skilled team with experience and expertise in real-time image processing. It’s very exciting.
“We’re spending a lot of time in China and the US West Coast engaging major software manufacturers to show off some of our technologies, such as making portrait mode photographs look a lot more seamless and real, and enhancing colours, depths and details by improving HDRs.”
Rhod was chief commercial officer at Swiftkey, spending three years here and three in Silicon Valley on the smart keyboard technology before it was bought by Microsoft. With experience selling to Apple, Google, Microsoft and major smartphone manufacturers, he is well placed to commercialise Spectral Edge’s technology.
“One particular area where the technology is strong is fusing an RGB image with infrared image in real-time. This can be for stills or videos. You pull in details from the infrared spectrum that you might not get in the colour spectrum.”
This is useful for surveillance cameras that use infrared at night, but could become an important feature for smartphone users too, cutting through mist or bringing out more detail in foliage, and providing stronger results on selfies.
“Infrared is more light and more data. The more data you can collect, the better the outcome,” says Rhod.
Spectral Edge’s first customer is NTT DATA Italia, which is incorporating its Vividteq technology into its broadcast infrastructure offering to help deliver sharper, crisper images with greater colour and contrast with reduced bitrate, regardless of set-top box or TV set limitations.
The smartphone market is the next big opportunity.
“There are different drivers with the smartphone people we work with,” says Rhod. “Some want to make their flagship better and unique and we can help. But also we can make the low-end look more like a mid-range. It’s about getting more out of what you have.”