Arm has released the Mali-C55, and said that it is its smallest and most configurable image signal processor to date.
Image signal processors (ISPs) are used in IoT vision systems such as commercial, industrial or domestic cameras and drones.
The ISP has multi-camera capability for up to eight separate inputs, support for image resolutions up to 8K and a maximum image size up to 48Mpixels. According to Arm, the Mali-C55 offers the most efficient combination of image quality, throughput, power consumption and silicon area.
Image quality has been enhanced, compared to earlier generations, including the Mali-C52, with improved tone mapping and spatial noise reduction, enhanced support for high dynamic range (HDR) sensors and seamless integration with machine learning accelerators for access neural networks for various de-noising techniques. Being able to send the output from the ISP directly to the ML accelerator reduces cost and processing time by having less data sent from device to cloud, explained Arm. This is achieved without having to compromise on inferencing, said the company.
In addition to multi-camera support and integration with ML accelerators, the Mali-C55 also includes industry standard AXI and AHB interfaces for easy integration with either Cortex-A or Cortex-M based SoCs.
By combining multiple Mali-C55 ISPs, larger image sizes can be achieved for applications that require greater than 48 MP capabilities, such as video conferencing.
The Mali-C55 occupies half the silicon area size of previous generations, significantly lower power consumption to extend battery life, said Arm. It also contributes to a reduction in the cost of devices, said the company.
The Mali-C55 works under a wide range of different lighting and weather conditions, reported Arm. Its performance and power budget mean it is suitable for power- constrained applications, such as smart camera and edge AI vision applications in surveillance and security as well as smart buildings. For example, said Arm, cameras will be able to detect more critical details, such as recognising precise information on car number plates as the vehicle is travelling at up to 75mph. They can also be used to capture higher resolution images inside and outside of buildings in security systems and in smart home hubs they can efficiently include advanced features like secure visual unlock.
There is also a software package for controlling the ISP, as well as tuning and calibration tools.