CUDA SDK for image processing on NVIDIA GPU

MRTech SK company proposes high performance, fast CUDA SDK for image and video processing on NVIDIA GPU. This SDK is a set of software components which correspond to image processing pipeline of any complexity for machine vision applications. The processing components and modules are done completely on GPU and this leads to real-time performance or above for the full pipeline.

MRTech CUDA SDK enables high-performance processing for both real time and offline application and is valuable when you have hardware with the dedicated NVIDIA GPU card or Jetson embedded module.


What this SDK works on:

  • Desktop and mobile NVIDIA GPUs including their top card GeForce RTX 2080 Ti, Titan RTX, GeForce RTX 2080 Super Max-Q, GeForce RTX 3090;
  • Workstation and Data Center NVIDIA GPU such as powerful card of Quadro and Tesla series;
  • Solution based on Tegra system-on-chip with Jetson Nano, Jetson TX2 and TX2i, Jetson Xavier NX, Jetson AGX Xavier modules.

MRTech CUDA SDK is proprietary cross-platform software implemented without using NPP library and focused primarily on performance and image quality issues. The SDK comes separately or as part of MRTech IFF when you need to build full image processing pipeline and to get the best possible performance and assured image quality for the whole system.

Special mention should be of GPU JPEG and JPEG2000 codecs which are part of the SDK. Their performance is much faster in comparison with CPU-based implementations of JPEG and J2K algorithms for image encoding and decoding.

An overview of the MRTech CUDA SDK components and what they allow to achieve is shown below.

Possible image sources:

  • High-resolution, high-performance, scientific cameras, Multi-camera system;
  • Image arrays stored on disk and other and other.


Processing tasks that can be effectively addressed:

  • High-quality debayers on CUDA;
  • Image denoiser library, resizer, many other algorithms;
  • Hardware / software compression and decompression;
  • Extra fast JPEG2000 encoding and decoding;
  • Custom GPU image processing.



Summary of CUDA SDK values


  • High Performance
  • Low latency
  • Better image quality
  • Reduced TCO for final applications
  • Compatibility with third-party SDKs on CPU and GPU
  • Ease of maintenance, updates and upgrades
  • Availability for full range of NVIDIA GPUs


The supply chain steps of this CUDA SDK are like those of MRTech IFF software:


  • Request and send your requirements;
  • Demo testing if necessary;
  • Mutual NDA and specification coordination;
  • Customization, testing when necessary;
  • Software delivery;
  • Consulting and supporting the software.


Please contact us for more information.