What’s So Great About EuclidVision?
With the proliferation of digital video in recent years in applications as diverse as cable TV live and on-demand broadcasting, Internet streaming, DVD and Blu-ray disc storage, and digital cinema with 4K resolution, the need for “better” video compression is stronger than ever. Working in this ultra-competitive field, EuclidIQ is often asked what makes its video compression technology different from that of any number of other video compression products and proposals on the market today.
The need to answer this question was the impetus behind the white paper, “EuclidVision: A New Approach to Model-Based Video Compression.” EuclidVisionTM is EuclidIQ’s video compression technology.
To see how EuclidVision produces compression improvements, it is important to understand first what the benefits – and limitations – of “standard” video compression methods are. To this end, the paper begins by reviewing the hybrid video coding schemes present in standard video codecs of the last 20+ years, in particular the block-based motion estimation and compensation (BBMEC) approach to temporal prediction.
To address the limitations of BBMEC, model-based compression schemes have been proposed through the years, most notably in the MPEG-4 Part 2 standard. However, these “traditional” model-based compression schemes have been hindered by implementation issues, especially the well-known challenge of correctly segmenting the video data.
In light of this, the paper then presents EuclidVision (EV), EuclidI Discoveries’ new approach to model-based compression that is able to reap the benefits of model-based processing without being encumbered by the implementation issues of previous model-based compression schemes. The paper details how EV does this, but from a bird’s-eye view, it is important to note two aspects of EV compression:
- EV applies modeling information indirectly to maintain compatibility with standard block-based video codecs, without having to implement segmentation and without having to encode feature/object shapes.
- Unlike many of the “latest and greatest” video compression schemes on the market, EV is not a pre-processing, post-processing, or filtering technique. EV improves one of the fundamental components of video encoding, the temporal prediction step.
To read the white paper, please click here.