Adaptive DCT/DST Modes for Lower-Volume Image Encoding

The patent applies parameter changes by transforming image data from the spatial domain to the frequency domain using transforms like DCT and DST. This changes the representation parameters of the image data, allowing for more efficient compression while maintaining quality. The transform coefficients are then quantized and encoded, achieving reduced data volume without significant loss of image quality.

The patent utilizes phase transitions in the sense of transitioning between different domain representations (spatial to frequency domain). This phase transition enables the image data to be represented in a form that is more suitable for compression, separating important frequency components from less important ones, thereby reducing data volume while preserving quality.

The patent applies segmentation by dividing the image into blocks or tiles, and further dividing these into smaller transform units. Each transform unit can be processed independently through the transform and quantization process. This segmentation enables more efficient encoding by allowing parallel processing and reducing the complexity of handling large images as a whole.

The patent implements dynamics by allowing adaptive selection of transform types (DCT, DST, or both) based on the characteristics of each transform unit. The encoder can dynamically choose the most appropriate transform for each block, optimizing encoding efficiency while managing complexity through localized adaptation rather than applying a single complex method uniformly.