Processor IP licensor ARM Holdings plc and Israeli startup EyeSight Mobile Technologies Ltd. have completed an optimization of EyeSight's gesture recognition software to run on ARM's Mali T600 series graphics processors.
EyeSight (Tel Aviv) was founded in 2005 and is a developer of gesture recognition and machine vision algorithms and software. The use of a GPU to perform gesture computation will result in increased efficiency and additional use cases, the company said.
The EyeSight software running on Mali provides access to a lexicon of gestures, hand signs and trackings. The use cases could include face and emotion detection and 3-D motion tracking for such things as finger pointing for selection.
"The optimization of gesture middleware solutions using Mali GPU-compute in combination with ARM Cortex-A processors using NEON technology is an industry first. It enables impressive performance, accuracy, robustness and efficiency," said Pete Hutton, general manager of ARM's media processing division, in a statement issued by EyeSight. NEON is the 128-bit single-instruction, multiple data (SIMD) extension the instruction set of Cortex-A series processors.
Gideon Shmuel, CEO of eyesight, commented: "By making it easier to integrate gesture into ARM powered devices, eyeSight and ARM are providing a route to really bring gesture to the masses."
EyeSight provides video pre-processing software that sits between the camera and processor and serves to prepare the image so that shapes of hands and movement of hands and fingers can be recognized. This helps with low-light conditions and low-resolution cameras.
The Eyesight technology allows Mali GPUs to process gestures in three dimensions when linked to stereoscopic, dual camera systems or devices with image and infrared depth perception.
While the use of gesture recognition software running on Mali GPUs could provide an efficiency boost for hand-waving control of mobiles, televisions and other consumer electronics it will require application developers to write their applications to EyeSight's software adding commercial complexity.
Meanwhile, Google has backed an alternative way to go, acquiring startup Bot Square Inc. of Palo Alto, Calif., for a sum reported to be close to $40 million. Bot Square's Flutter software works directly with a single webcam image sensor. The company is currently selling Flutter as a downloadable app for use with Mac and Windows notebook computers where it can be used to control such applications as iTunes, Spotify, Netflix, Youtube.