Image Coding TRIZ Cases

Deep neural network filters compensate quantization errors during decoding, improving image quality without excessive computational overhead.

Adaptive collocated block selection reduces memory bandwidth and implementation complexity during video decoding.

A compression system selects between planar and delta color schemes based on data plane counts to reduce transfer volume.

Encoded update parameters modify a deep neural network decoder, resolving the trade-off between objective fidelity and perceptual visual quality.

Replacing missing shading atlas blocks with stored prior versions reduces latency and packet loss impact while preserving rendering accuracy.

Encoding apparatus compresses point cloud data by representing the sensing path as a two-dimensional curve.

A long-term palette system signals color information at a higher level to enable dynamic construction and updating of palettes using neighboring block data.

Viewport-dependent DASH adaptation sets optimize bandwidth by delivering high-quality content only within the active viewing region.

A point cloud data processing device segments geometry into patches for parallel encoding and decoding operations.

A point cloud coding method selects attribute inter prediction based on rate and distortion information.

Neural networks compress video frames into quantized latent representations, reducing data quantity while maintaining acceptable image quality.

Encoder projects point clouds into patch images with quantized depth levels, reducing data volume while maintaining spatial integrity.

A signal-based encoder converts point cloud data into compressed waveforms for network transmission.

Texture maps modify digital images to match item shapes, resolving rendering inaccuracies for complex geometries and textured finishes.

Split scanning adapts pixel order to block texture, resolving fixed-scan compression limits.

Coordinates texture and geometry decoding to prevent frame mismatches, ensuring consistent image quality during volumetric video playback.

Embedding codec type identifiers in file control information allows devices to quickly determine data types, reducing processing loads and increasing speed.

Segmenting residual vectors into modulus-based sub-codebooks reduces redundant codewords, saving storage space while maintaining encoding precision.

Segmenting reference pixels reduces processing cost while maintaining prediction accuracy in HEVC LM mode.

Encoding video frames with mesh and texture data enables interactive viewpoint exploration while maintaining production efficiency.

Segmenting high dynamic range depth buffers into tiles minimizes quantization errors and reduces transmission latency by over 1000 times.

Replacing PReLU with LeakyReLU fixes training divergence and improves video coding performance.

A video compression method exploits correlations between color components to increase data reduction factors while maintaining structural similarity.

Inferred direct coding mode applied to isolated nodes reduces processing complexity in point cloud tree structures.

Processing quantized discrete cosine transform coefficients embeds watermark data without increasing bitrate or introducing perceptible artifacts.

Encoder calculates priority values from error metrics and encoding costs to select edges for collapse, balancing mesh quality with compression efficiency.

A 3D mesh compression method clusters repeating geometric components to reduce data size and accelerate decoding.

A video coding apparatus selects partitioning patterns from a predetermined group based on segmentation mask sample values to associate the correct pattern with coding blocks.

Groups sorted point cloud data by spatial correlation to improve decorrelation efficiency and coding performance.

N-ary tree nodes encode using inter and intra prediction modes to reduce data volume while maintaining coding efficiency for large point cloud datasets.

Layered V-PCC encoding reduces latency by allowing receivers to process essential geometric information first.