Khronos Finalizes and Releases OpenVX 1.0 Specification for Computer Vision Acceleration
Fast, portable, power-efficient vision processing across diverse hardware architectures; Full conformance test suite and Adopters Program immediately available;
Khronos to ship open source OpenVX implementation by end 2014
October 20th, 2014, – The Khronos™ Group today announced the ratification and public release of the finalized OpenVX™ 1.0 specification, an open, royalty-free standard for cross platform acceleration of computer vision applications. OpenVX enables performance and power-optimized computer vision processing, especially important in embedded and real-time uses cases such as face, body and gesture tracking, smart video surveillance, advanced driver assistance systems (ADAS), object and scene reconstruction, augmented reality, visual inspection, robotics and more. In addition to the OpenVX specification, Khronos has developed a full set of conformance tests and an Adopters Program, that enables implementers to test their implementations and use the OpenVX trademark if conformant. Khronos plans to ship an open source, fully-conformant CPU-based implementation of OpenVX 1.0 before the end of 2014. The full OpenVX 1.0 specification and details about the OpenVX Adopters Program are available at www.khronos.org/openvx.
OpenVX defines a higher level of abstraction for execution and memory models than compute frameworks such as OpenCL™, enabling significant implementation innovation and efficient execution on a wide range of architectures while maintaining a consistent vision acceleration API for application portability. An OpenVX developer expresses a connected graph of vision nodes that an implementer can execute and optimize through a wide variety of techniques such as: acceleration on CPUs, GPUs, DSPs or dedicated hardware, compiler optimizations, node coalescing, and tiled execution to keep sections of processed images in local memories. This architectural agility enables OpenVX applications on a diversity of systems optimized for different levels of power and performance, including very battery-sensitive, vision-enabled, wearable displays.
“Increasingly powerful and efficient processors and image sensors are enabling engineers to incorporate visual intelligence into a wide range of systems and applications,” said Jeff Bier, founder of the Embedded Vision Alliance. “A key challenge for engineers is efficiently mapping complex algorithms onto the processor best suited to the application. OpenVX is an important step towards easing this challenge.”
The precisely defined specification and conformance tests for OpenVX make it ideal for deployment in production systems, where cross-vendor consistency and reliability are essential. OpenVX is complementary to the popular OpenCV open source vision library that is also used for application prototyping but is not so tightly defined and lacks OpenVX graph optimizations. Khronos has defined the VXU™ utility library to enable developers to call individual OpenVX nodes as standalone functions for efficient code migration from traditional vision libraries such as OpenCV. Finally, as any Khronos specification, OpenVX is extensible to enable nodes to be defined and deployed to meet customer needs, ahead of being integrated into the core specification.
Industry Support
“AMD is an enthusiastic advocate for natural user interfaces enabled by computer vision,” said Greg Stoner, senior director of application engineering, Heterogeneous Applications and Solutions, AMD. “As a proponent of open standards and a provider of highly-parallelized architectures ideal for computer vision, we embrace the OpenVX standard and look forward to standardized proliferation of these experiences throughout the industry.”
“OpenVX will be a great tool for future development of computer vision applications,” said Johan Paulsson, CTO at Axis Communications. “We see excellent potential for making development efforts future-proof by separating algorithms from hardware.”
“Cadence is integrating OpenVX into our Tensilica® Imaging/Vision Library software development kit to enable higher performance and power optimization for our scalable and configurable IVP-EP DSP cores, which are widely adopted for imaging, computer vision and automotive drivers’ assistance applications,” stated Dr. Chris Rowen, CTO of the IP Group, Cadence.
“CEVA applauds the Khronos Group on achieving this important milestone with the release of OpenVX specification,” said Eran Briman, vice president of marketing at CEVA. “With an extensive list of vendors currently sampling their computer vision ICs based on the CEVA-MM3101 DSP and a growing developers community, it is clear that the optimal way to meet the stringent performance and power requirements of computer vision applications is to offload CV processing to a dedicated processor. OpenVX enables this for developers in a seamless manner and we are now integrating it into our CEVA-MM3101 Application Developers Kit (ADK).”
“As one of the first innovators in embedded vision, CogniVue will be supporting a compliant OpenVX implementation for our APEX Image Cognition Processing technology with announcements coming,” said Tom Wilson vice president of product management and marketing at CogniVue.
“We see OpenVX as a promising tailor-made standard tool for developing performant, low-power, while portable computer vision applications for our mobile devices,” said Zhouhong, president of
Central Hardware Engineering Institute, Huawei. “Based on its tightly defined nature, we also see
OpenVX to serve as a standard benchmark suite which can promote the development of ever-more power efficient computer vision accelerators, a component likely to become must-have for coming mobile products.”
“Vision applications will be a defining feature and differentiator for products in a wide range of markets including mobile, smart home, automotive, retail analytics, public safety and more. Imagination’s IP cores including PowerVR GPUs and video processors, Raptor imaging processors and MIPS CPUs are already used in vision applications, and we expect this trend to accelerate,” said Peter McGuinness, director of Multimedia Technology Marketing, Imagination. “We are delighted to be involved with defining the specification for OpenVX 1.0, which we believe is a valuable starting point for accelerating creation and adoption of vision applications.”
“Intel supports and welcomes the adoption of OpenVX as an important element in proliferating computer vision usage models,” said Ofri Wechsler, Intel Fellow. “In fact, Intel has a long history in supporting the development of the computer vision domain spanning the creation of OpenCV more than two decades ago to the active optimization of OpenCV 3.0 for Intel CPUs and Intel Processor Graphics”
“OpenVX enables disrupting computer vision applications that run on low-power mobile and wearable devices,” said Victor Erukhimov, CEO of Itseez, Inc., and chair of the OpenVX working group. “We see it as an optimization layer for many of our projects, from sophisticated middleware for smartphones to advanced driver assistance technologies and mobile 3D scanning.”
“We are excited to contribute to the next generation of computer vision applications,” says A.G.K. Karunakaran, CEO of MulticoreWare. “Our experts are already working with leading device OEMs and semiconductor companies on the development of OpenVX platforms and computer vision libraries that leverage multicore heterogeneous processors including CPUs, GPUs, DSPs, VSPs, and other programmable architectures.”
“Real-time vision processing combined with advanced graphics is inter-twining the real and virtual worlds to enable true visual computing applications,” said Neil Trevett, president of the Khronos Group and vice president of mobile ecosystem at NVIDIA. “NVIDIA is integrating OpenVX into our VisionWorks SDK so GPU-accelerated vision nodes can be easily combined into pipelines for a range of advanced vision applications.”
“Renesas is pleased to take part in the launch of OpenVX that is expected to have a significant impact on the vision application domain,” said Masayuki Mizuno, vice president, chief of Incubation Center at Renesas Electronics Corporation. “We believe OpenVX will break new ground in vision applications by introducing graph-based optimization as well as by defining standard vision APIs.”
“Samsung is committed to bringing novel, useful, and fun computer vision applications to our mobile devices. The standard framework of OpenVX lets us develop portable applications for diverse application processors and consumer devices, from tablets and handsets to wearables,” said Mike Polley, senior vice president and head of the Samsung Mobile Processor Innovation Lab. “Samsung encourages the industry to support OpenVX to enable broad availability of computer vision on mobile platforms, allowing Samsung and independent software developers to unlock the potential of our advanced processing architectures and deploy pervasive, revolutionary and efficient computer vision in power-constrained environments.”
“Texas Instruments is committed to enabling advanced vision analytics processing and to support customers to streamline their embedded development,” said Jason Jacob, processor general manager — ADAS, Texas Instruments. “As a member and contributor of the Khronos OpenVX working group and efforts, TI believes the release of the OpenVX specification is a key milestone in enabling Open Standards based embedded compute frameworks to drive a faster adoption rate in the market. Our digital signal processors (DSPs) provide the performance lift and power efficiency for vision analytics and TI will be enabling OpenVX on these architectures.”
“videantis congratulates the OpenVX team on reaching another major milestone,” said Hans-Joachim Stolberg, CEO at videantis. “We see OpenVX as a key open standard that enables efficient acceleration of computer vision algorithms, driving new applications such as automotive driver assistance systems, always-on camera applications, gesture interfaces, 3D capture, and augmented reality. We’re proud to bring support for this new standard to our v‑MP4000HDX scalable unified video/vision processor architecture.”
“Vivante is working with leading industry partners to deliver high performance, real time, intelligent vision processing solutions in mobile, automotive and smart security products through our OpenVX based GC7000 VX Series GPUs,” said Wei-Jin Dai, CEO and President of Vivante. “This innovative design enables SoC vendors to use the same processing core for embedded vision acceleration and photorealistic 3D rendering, while keeping power consumption within mobile levels. Through a dynamic VLIW vision instruction set and enhanced shader extensions to achieve single cycle API efficiency, this new GPU Vision approach will allow rapid adoption of vision applications on Vivante OpenVX platforms.”
About The Khronos Group
The Khronos Group is an industry consortium creating open standards to enable the authoring and acceleration of parallel computing, graphics, vision, sensor processing and dynamic media on a wide variety of platforms and devices. Khronos standards include OpenGL®, OpenGL® ES, WebGL™, OpenCL™, SPIR™, SYCL™, WebCL™, OpenVX™, OpenMAX™, OpenVG™, OpenSL ES™, StreamInput™, COLLADA™ and glTF™. All Khronos members are enabled to contribute to the development of Khronos specifications, are empowered to vote at various stages before public deployment, and are able to accelerate the delivery of their cutting-edge media platforms and applications through early access to specification drafts and conformance tests. More information is available at www.khronos.org.
###
Khronos, DevU, StreamInput, SPIR, SYCL, WebGL, WebCL, COLLADA, OpenKODE, OpenVG, OpenVX, VXU, glTF, OpenSL ES and OpenMAX are trademarks of the Khronos Group Inc. ASTC is a trademark of ARM Holdings PLC, OpenCL is a trademark of Apple Inc. and OpenGL is a registered trademark and the OpenGL ES and OpenGL SC logos are trademarks of Silicon Graphics International used under license by Khronos. All other product names, trademarks, and/or company names are used solely for identification and belong to their respective owners.