Segmenting screen content into distinct regions reduces CPU-GPU bandwidth costs while maintaining high compression ratios.
A video compression system partitions input streams to optimize spatial and temporal resolution tradeoffs for efficient data representation.
A video preprocessor bypasses SKIP macroblock processing to reduce computational load.
Processes base and dependent views to resolve transcoding quality issues in single-view displays.
A filtering system adjusts intensity based on local image characteristics to reduce visual artifacts in compressed video streams.
A function-based block matching technique generates representative values for video blocks to accelerate encoding processes.
Mapping motion vectors between non-uniform and high-accuracy uniform grids resolves scaling inaccuracies caused by filter restrictions in video coding.
A region-based adaptive loop filtering approach divides video frames into blocks to reduce computational complexity and power consumption.
Adaptive reference filtering compensates for sharpness mismatches between frames, improving prediction accuracy and coding efficiency.
Segmented processing reduces redundant calculations during fractional pixel expansion, improving motion estimation precision and speed.
A data compression system processes images using 2x2 blocks with differential pulse code modulation and pulse code modulation modes.
A video encoding apparatus selects reference pixels within a predetermined distance range to generate predicted signals for target blocks.
Adaptive chrominance downsampling adjusts ratios based on frame characteristics to resolve the trade-off between code amount and subjective image quality.
INTRA_DIRECT method inherits intra prediction directions from base layer to enhancement layer macroblocks.
Predominant motion detection generates global vectors through histogram formation and filtering, capturing fast global motion scenes that fixed sampling misses.
Distributed image processing system estimates network bandwidth to adjust compression levels.
A hybrid reference construction unit builds textures combining frame and field data to decode interlaced macroblocks accurately.
Predictively encoding filter coefficients using horizontal and vertical symmetry reduces transmitted data volume, resolving compression efficiency trade-offs.
Estimating macroblock sizes early prevents re-processing overhead and wasted hardware cycles by making slice cap decisions prior to entropy coding.
A blurring process replaces pixels on blocking boundaries with randomly selected pixels within predefined regions.
Subdividing video frames into uniform pixel blocks allows selective encoding of changed regions, reducing decoding delay on low-performance devices.
A transcoding system modifies video stream spatial resolution during decoding to optimize quality for client devices.
Reduced-scale motion search modules refine motion vectors via downscaled pictures to lower computational complexity while maintaining high-resolution accuracy.