Augmented & Virtual Reality

MulticoreWare has been working with Augmented and Virtual Reality startups, hardware and silicon vendors, and SDK, library and app developers to enable high-performance image processing and vision algorithms on their platforms.  We have worked with customers on their existing products to help upgrade software capability and efficiency, and have also contributed to hardware design choices on future products.

Power Efficiency

Porting vision and imaging algorithms while retaining power efficiency is vital for battery powered headsets and devices. Talk to us if you have tight power constraints but still need quality vision capabilities.

High Precision Vision

Our team of vision experts has experience in a wide variety of algorithms and use cases.  We can help improve the quality of your vision algorithms while working within your sensor and hardware constraints.

Performance Optimization

We’ve optimized vision and imaging algorithms for GPUs, FPGAs, DSPs, custom embedded chips, and more.  We can leverage our expertise to help you optimize your algorithms for any target platform, big or small.

Solutions for Multiple Tiers

Semi-conductor & Hardware Vendors

  • Hand-tuned assembly code development for your platforms.
  • OpenVX, OpenCL, Renderscript stack development and conformance.
  • Optimization of algorithms/libraries for your architecture.
  • Video camera driver development, camera integration and correction.

SDK and Software Developers

  • Optimization of libraries for target hardware architectures (mobile & desktop).
  • Acceleration of code for multiple platforms - Cloud, GPU, DSP, FPGA, embedded, etc.
  • Custom algorithm design, testing, and optimization for target hardware.
  • OpenVX, OpenCV, or custom library development consulting and services.

AR/VR Service Providers

  • Software and hardware integration consulting & services.
  • Custom algorithm performance testing, optimization, and modification.
  • Hardware design choice consultation, benchmarking and comparison.
  • Latency reduction and optimization for multiple cameras and sensor fusion.

Computational Photography & Image Post-Processing

Mobile, video, automotive, defense and other imaging and vision applications demand fast, high-quality processing with a low power and memory footprint.  Our experience in designing and optimizing imaging and vision libraries and functions provides us with the tools and expertise to streamline your image processing pipeline.

Sample Algorithms

Image Stitching for Panoramic and Surround View

Image stitching is a computationally intensive task required by nearly all smartphones and surround view cameras.  MulticoreWare has accelerated and improved stitching quality for mobile, automotive, and defense applications.

Optimized High Dynamic Range

Optimizing High Dynamic Range (HDR) can significantly improve the performance and power efficiency of combining multiple images and improve user experience.  We have also applied HDR optimizations for video applications.

Motion Blur Reduction

Optimizing High Dynamic Range (HDR) can significantly improve the performance and power efficiency of recombined HDR images and improve user experience.

Libraries & Standards

MulticoreWare has been involved in the design and development of the OpenCV open-source vision library and the Khronos OpenVX standard for portable and power-efficient vision processing.  We are well-versed in implementing and using these libraries and standards and can help get your OpenCV or OpenVX apps built and optimized for your target platform.

MulticoreWare is one of the leading developers of the OpenCV computer vision library.  Our development teams have accelerated dozens of OpenCV kernels using OpenCL, contributing these improvements back to the OpenCV project.
OpenVX is the latest standard for power-efficient and portable computing.  MulticoreWare has been involved in developing the standard since v1.0 and is experienced in implementing and designing OpenVX applications.

Imaging & Vision for Any Platform

Enable state-of-the-art image processing and vision algorithms on any platform.  We’ve ported OpenCV, OpenVX, and custom algorithms to dozens of architectures.

Embedded

We can target embedded designs found in modern AR/VR headsets, cameras, vehicles, and phones.  Improve your product by leveraging our experience in creating vision libraries for power-efficient embedded devices.  We’ve developed for devices with a variety of requirements, including stereo-vision, low-latency feature detection, and more.

Mobile

Mobile devices and handsets equipped with cameras represent the largest market for image and vision processing.  We’ve worked with device manufacturers, app developers, and semiconductor vendors to improve the capability of their devices by utilizing the CPU, GPU, DSP, and other processing elements available in modern SoCs.

Workstation & Cloud

For ultra-high quality and intensive applications, GPU or FPGA enabled workstations and cloud servers can provide the computational power necessary to handle large scale tasks and video processing.  We can help design and port your algorithms to fully utilize the power of modern GPUs, FPGAs and server-class hardware.

We've ported image processing and vision algorithms on these platforms and more.