Super Resolution Enabled CODEC

Dimension, Inc. has created a disruptive video CODEC technology that exchanges superresolution reconstruction for a reduced transmission bandwidth requirement. In operation, raw video frames are first profiled for superresolution processing and then downsampled. Downsampled frames are then compressed using a second-layer CODEC, (e.g. MPEG-4, H.264, or MJPEG), and subsequently merged with profile data for network transmission. At the receiver, the process is reversed; (i) superresolution profile data is first extracted, (ii) downsampled video is upsampled, and finally (iii) superresolution profile data is applied to frame reconstruction at the original resolution. The unique aspect of this two-layer Superresolution-Enabled CODEC (SRECODEC) is superresolution and existing lossy CODECS work together in extending bandwidth efficiency with quality to levels previously unattainable with just lossy CODEC alone.

what we do

Dimension's software technology turns ordinary SD & HD video into beautiful 4K & 8K footage!

Dimension has developed a one-of-a-kind core technology based on its patented Photometric Warp image reconstruction technology, introducing a quality of video and speed of transfer not seen before in the video technical arena.

Dimension's latest breakthrough resets the industry standards with HD-to-4K/120fps and HD-to-8k/90fps at less than single frame latency.

Applications Include:
  • Digital Cinema
  • HD → 2K → 4K → 8K and beyond!
  • VR/AR, Streaming Video, and Video Post-Production
Our technology provides:
  • Highest quality results
  • Cinematic image quality
  • Sharpness and details
  • HD-to-4k/120fps and HD-to-8k/90fps < 1 frame latency Real-Time Video Streaming
  • GPU, Multicore, and FPGA implementations

Soon to be available as a Cloud-based SaaS (Software as a Service) application!

press releases

how we do it

Dimension upscaler technology incorporates a proprietary (patent pending) Photometric Warp (‘p-Warp’) superresolution technique affording critical advantages in terms of: (1) reconstruction processing gain and (2) algorithmic/architectural scalability. These advantages accrue as a direct consequence of p-Warp superresolution formulation in terms of nonlinear image processing mathematics. As such, nonlinear image processing enables efficient and accurate reconstruction at a reduced algorithmic complexity when compared to alternative design approaches such as Machine-Learning and AI. Taken together, all advantages combine to make AR/VR and real-time streaming of upscaled video content a practical reality at ‘HD-to-4k’, ‘HD-to-8k’, and beyond.

why you need it

Dimension’s Photometric Warp Superresolution technology offers the highest real-time performance of any upscaler available in the marketplace. In terms of efficiency, throughput, and quality, Dimension remains unsurpassed by any other streaming-video solution.

View or download our videos here: Video Demos

demos

Dimension's Story

Dimensions’s core products are built on representing images, particularly video, using a mathematical description that is both efficient and not tied to a specific resolution (e.g., screen width or height). Dimension uses this type of representation to accomplish image compression and also increase the number of pixels of the images (e.g., from SDTV to HD, HD to 4K, 4K to 8K, and beyond). While these are just a few of the use cases that Dimension’s technology makes possible, in the new world of artificial intelligence (AI), the company can enhance many aspects of existing machine learning systems.

When Dimension transitioned from its fractal technology to its Superresolution-Enabled Video CODEC (SREC), it gained two distinct advantages. It increased quality for a given size of video representation, but whereas fractals were highly asymmetric and required the ability to process video offline, SREC is implemented in real-time, which means it can be interpolated into many workflows without negatively impacting latency.

Compression and upscaling are only a few of the potential uses for Dimension’s technology. With its latest technology, data that describes the image can be easily separated from noise, which is an important first step in machine learning (ML) classification pipelines (e.g., AI systems). Because Dimension’s SREC runs in real-time, it can easily fit into many existing workflows better and more efficiently.

With a deeper integration into a ML pipeline, Dimension can go even further. The first step of many AI systems is to embed a description of an image into a more meaningful representation, at least for the computer. The technology can be used as a preprocessor to speed that process, as the technology can already extract visually meaningful features. Not only will this potentially improve the performance of existing systems, but it can use this new representation as a query language to ask for images that "look like" a provided example.

Dimension is now pursuing patents on many of these new aspects of its technology, which will generate an exciting and new set of potential customers, especially in the emergent BigData market sector.

Latest News

Dimension Inc is well known for development of advanced image processing technologies pertinent to high-performance video applications such as IoT, Surveillance, Analytics, Video Conferencing, VR/AR, and Streaming Video Platforms. In particular, Dimension has focused upon image and video reconstruction, denoise, and compression (CODEC) as those processing components most critical to video system performance optimization. From a technical perspective, this work has been successful in creation of new algorithms of extraordinary power.

View or download our Big Data Whitepaper here: Big Data Whitepaper

company

Dimension, with its growing patent portfolio, many years of R&D experience, and now with the introduction of our advanced ‘smartTiler’ super-resolution technology, is fully prepared to assist our customers in achieving a higher level of quality and speed in their critical video and image applications:

  • Virtual Reality/Augmented Reality
  • Next Generation Television (4k, 8k, and beyond)
  • Streaming Internet Video
  • Thermal Imaging
  • Video Surveillance
  • Medical Imaging
  • Video Pre-Conditioning for Machine Learning Application

Demonstrations and solutions are available today.
to contact us!

team

Copyright © 2008-2024 Dimension, Inc. All Rights Reserved.