Abstract: Using image segmentation in video compression brings about a limitation in video quality when a segment moves in position from frame to frame. The limitation arises because color contributions (bleeding or blurring) naturally occur between neighboring segments. The above-identified problem is overcome by providing solutions to compensate for color contributions between neighboring segments. In accordance with one embodiment, a method and apparatus de-blurs an image segment. De-blurring involves removing from the segment approximate color contributions from neighboring segments. This results in a segment that is approximately independent of color contributions from neighboring segments. In accordance with another embodiment, a method and apparatus re-blurs an image segment. Re-blurring involves adding to the segment approximate color contributions from a new arrangement of neighboring segments. This results in more realistic rendering of the segment, as it is located in the new arrangement.

Abstract: In one embodiment, an image is broken up into multiple regions or segments, where each segment may be of arbitrary shape, and a transform (multi-scale or otherwise) is applied on the set of segments. In another embodiment, pattern adaptive prediction is used when predicting the next finer level of the transform pyramid. The pattern adaptive prediction uses the parent grid to determine what geometry of a filter is to be used when predicting the child grid. At the boundaries of the domain, the pattern adaptive prediction can be coupled with the domain adaptive prediction technique.

Abstract: In accordance with an embodiment, a method of encoding includes generating for each transform point a double difference coefficient (comprising the difference between a modeled difference coefficient and a raw difference coefficient) and encoding as an adaptive difference coefficient for each transform point either the double difference coefficient or the raw difference coefficient. Whether the double difference coefficient or the raw difference coefficient is selected to be the adaptive difference coefficient depends on which one provides more efficient coding. A method of decoding includes receiving the adaptive difference coefficients from the encoder, applying the same modeling and transform as the encoder to generate the modeled difference coefficients, generating corrective difference coefficients (from the adaptive difference coefficients and the modeled coefficients), and inverse transformation using the corrective difference coefficients.

Abstract: In one embodiment according to the present invention, relative z-ordering of segments in a digital image is determined. A method comprises forward and backward motion matching of image regions to determine overlap, followed by the creation of relationships (e.g., pairwise relationships) between regions and comparing the result with the original image to determine the relative z-ordering.