Author(s): Raghu Gopal
This week, UK chip designer ARM announced the acquisition of Apical, a developer of chip-based vision systems that provide smartphones and computers with the ability to react to contextual stimuli. We expect that ARM will incorporate Apical's technology into future microprocessor and system-on-chip designs.
According to ARM CEO Simon Segars, Apical's technology will provide new opportunities for designers of autonomous vehicles and security systems, as well as other connected devices. Computer vision is in its early stages and Apical is certainly one of the companies at the forefront.
Apical's technology is already used in 1.5 billion smartphones and about 300 million other consumer and industrial devices such as IP cameras, digital cameras, tablets and smartphones.
Apical develops imaging technology that allows electronic devices to learn from their environments and act on the most relevant information by employing intelligent processing. For example, its technology can help a security camera distinguish between a human and an animal. Assertive Camera technology is another of Apical's developments. It offers a range of software packages and silicon-based image signal processors for reducing image noise, managing colour and shooting high dynamic range images. Apical Resident Technology uses a smartphone camera sensor, plus image analysis and machine-learning software in a local device, to build up profiles of people in a household through gesture, gait and route analysis without recording any video. It then reports when it sees someone it does not recognize such as an intruder.
ARM designs chips that other firms can manufacture, or license to incorporate into their chipsets. In addition to its Cortex family of microprocessors, ARM also licenses Mali graphics cores. It now plans to add Apical's Spirit blocks to its portfolio. These silicon blocks process raw sensor data or video into a machine-readable representation of an image in an energy-efficient way. ARM and its customers could then use them to add forms of computer vision to future low-power devices.
Like many of its competitors and clients, ARM sees one of its biggest opportunities in the Internet of things. In particular, the company envisions growing awareness, and connectivity being built into automobiles as well as even the most mundane household items such as light bulbs. All these require very tiny chips that can process information and send the data online. ARM aims to develop and license the chip blueprints for these just as it does for most smartphones.
As the growth in global smartphone shipments slows, ARM is looking at various options to diversify its portfolio. We expect it to venture into new areas such as smart vehicles, security systems, robotics, as well as industrial and retail applications where advanced image processing could be deployed. ARM is not the only silicon design company looking to diversify its portfolio. Other companies like Intel, Nvidia and Qualcomm are also gearing up to exploit opportunities presented by the Internet of things, and, in the process, develop chips optimized for computer vision and other forms of machine learning within devices.