Many times, progress isn’t possible without sacrifice, without compromising. But as we found out today at Mobile World Congress, there are some companies who just make it work. As we talked with the guys from global leader in computational imaging and computer vision FotoNation, an Xperi brand, we realized that their latest product is capable of a real feat – decrypting neural networks in real time, without penalty to individual performance #machinemagic
During one of the stops in Hall 1 of the Fira Gran Via, we found out about an industry first – the third-gen IPU from FotoNation. This is the abbreviation for Image Processing Unit, the company’s imaging engine. This time around, it’s AI-enabled, meaning it “provides secure edge computing by decrypting neural networks in real time with no penalty on performance.”
Why is this so vital? Well, at this moment in time, most similar products can’t deliver super high quality camera experiences without compromising bandwidth and power consumption. FotoNation’s 3-gen IPU, though, is able to run various, simultaneous image enhancement solutions without sacrificing the performance of each of them.
The fact is, we’re entering an Edge of Things era, where our priorities will shift a bit, as well as our focus when it comes to cameras: “We believe that AI and neural networks will play a significant role in the future of real-time camera applications on the edge, where latency, lower bandwidth and substantially reduced power consumption are key in delivering always-on applications,” said Geir Skaaden, executive vice president and chief products and services officer, Xperi Corporation.
When that happens, companies that took the lead in the industry prior to that moment will have the upper hand. What we can’t be more happy about is the attention machine learning is receiving and the progress that it is made so that, us, consumers can enjoy the best possible experiences.