A new vision processor that enables scalable edge AI performance in smart camera applications

     Texas Instruments has introduced a new family of six Arm® Cortex® based vision processors that enable designers to add more vision and artificial intelligence (AI) processing capabilities at lower cost and more energy efficiency in applications such as visual doorbells, machine vision and autonomous mobile robots. These processors are built on system-on-chip (SoC) architectures that enable extensive integration. Integrated components include Arm Cortex-A53 or Cortex-A72 central processing units, third-generation Texas Instruments image signal processors, internal memory, interfaces and hardware accelerators that provide AI processing capabilities ranging from 1 trillion to 32 trillion operations per second (TOPS) for deep learning algorithms.

     The new family of AM62A, AM68A, and AM69A processors is supported by open source evaluation and model development tools as well as common software that can be programmed through industry-common application programming interfaces (apis), frameworks, and models. With this platform of vision processors, software, and tools, designers can easily develop and extend edge AI designs across multiple systems, while reducing time to market. Ti's new vision processor brings intelligence from the cloud to the real world by removing cost and design complexity barriers when implementing visual processing and deep learning capabilities in low-power edge AI applications.

     "Real-time response in electronic devices, where the world is evolving, requires local decision-making and improved power efficiency," said Sameer Wasson, vice president of Texas Instruments' processor division. This new processor family of affordable, highly integrated SoCs enables the future of embedded AI by enabling the use of more cameras and increased visual processing in edge applications."