The proliferation of high-bandwidth networks and advanced display technologies is fueling an exponential rise in video data consumption. Industries from entertainment to enterprise are grappling with escalating storage and transmission costs, alongside user expectations for pristine, low-latency content. Regulatory pushes for energy efficiency also favor technologies that reduce data footprint. This patent offers a timely solution to these converging market forces, enabling more sustainable and performant digital infrastructure.
Increases prediction accuracy by up to 20%, significantly reducing data volume.
Enables high-quality, low-latency streaming, maximizing user experience.
Secures a unique algorithm, differentiating from competitors.
The patent protects a specific method for dynamically determining weighting coefficients in prediction block generation based on the intra-prediction mode of surrounding blocks. The successful grant of this patent, despite two office actions, demonstrates its robust structure and strategic IP management, establishing clear inventiveness over conventional fixed weighting or simple prediction methods.
This patent focuses on prediction block generation. White space could exist in post-processing enhancements, adaptive bitrate streaming algorithms, or hardware-specific encoding accelerators not directly tied to the prediction mode determination.
Assuming a 15% reduction in data transmission for video streaming services. For example, with a monthly data transmission of 5PB (5,000TB) and a communication cost of $13.33/TB (AI est.), the annual communication cost reduction could be calculated as (5,000TB × $13.33/TB × 0.15 × 12 months) = ~$120K (AI est.).
X: Encoding Efficiency (Data Reduction)
Y: Image Quality Retention (Visual Fidelity)