Intel Software has unveiled a 2020.1 version of its OpenVINO developer toolkit for AI and machine learning implementations with human vision emulation.
According to Intel, its AI and Computer Vision ecosystem has now gone beyond the theoretical to real-life implementation – ripe for systems integrators and others to work with.
“They’ve been shown to work in practice,” Intel said in its official documentation. “Momentum is building across almost every industry to use AI.
“Those who continue to resist or lag behind now run a very real risk of being overtaken by competitors who are more inclined to innovate.”
Intel’s Open Visual Inference and Neural network Optimization (OpenVINO) toolkit aims to provides developers looking to develop cost-effective real-time vision applications with improved neural network performance on a variety of Intel processors.
“The toolkit enables deep learning inference and easy heterogeneous execution across multiple Intel platforms (CPU, Intel Processor Graphics), providing implementations across cloud architectures to edge devices,” according to Intel.
“This open source distribution provides flexibility and availability to the developer community to innovate deep learning and AI solutions.”
The company has also launched a guide including learnings and best practices from Common Vision implementations so far, which can be downloaded here.
“There are learnings and best practices to be taken from stories like these and applied to your own projects,” says Intel. “The OpenVINO toolkit is designed to help overcome some of the common challenges inherent in developing more intelligent vision software.”
The toolkit includes a range of pretrained models for common computer vision use cases, from human pose estimation, gaze estimation and emotion recognition to pattern and deviation recognition and ‘smart classroom’ use cases, it said.
Full notes on the 3.2 release, which extends support and optimisation for workloads across Intel hardware including accelerators and enables deep learning inferences from edge to cloud, are available here.
Intel OpenVINO solutions are being promoted to AI researchers, application developers, data scientists, and any other customer seeking high-performance deep learning, streamlined development and workflows and cross-platform scalability across Intel architectures.
Markets include life sciences and healthcare, smart cities and Internet of Things applications, Industrial Internet of Things (Industry 4.0), and retail.
According to Intel, the overall computer vision market is expected to reach $17.4 billion and the video analytics market $11.2 billion in about three years. Deep learning is forecast to create $13 trillion in additional economic activity over the next 10 years or so, it said.
