Abstract: Processes for compressing images with sparse histograms are disclosed. The image is divided into blocks, and a bit budget is assigned for each block. The pixels of a block are converted and coded bit-plane by bit-plane, starting from the Most Significant Bit (MSB) and going towards the Least Significant Bit (LSB). The pixels of the block are partitioned into groups. Each group contains pixels that have same value. Moving from the MSB to the LSB, the groups in each bit-plane are processed. When processing a group, the encoder sends a “0” if all group members have same bit value at the current bit-plane being processed, followed by the bit value; otherwise, the encoder sends a “1”, followed by refinement bits for each pixel of the group, and the encoder splits the group.

Abstract: Dual-mode compression of images/videos for reliable real-time transmission includes a scalable block-based video compression. The method provides graceful quality degradations in case of channel bandwidth changes. Prediction for encoding is avoided, and the process is less complex with less memory requirements. The method involves processing each block independently with the advantage that errors are not able to propagate beyond current block boundaries. Two different encoding modes are used, natural image coding for acquired images and graphics coding for computer generated images. The proper encoding mode is chosen by comparing performance of the two modes for a particular block to be encoded. An embedded bitstream is generated by the encoder in accordance to the available channel bandwidth.

Abstract: Method and apparatus for hierarchical bi-directional intra-prediction in a video encoder is described. In some examples, initial reference blocks are selected in an image. Sets of blocks in the image are predicted successively over a plurality of stages based on a respectively plurality of templates, each of the plurality of templates including reference blocks selected from at least one of: the initial reference blocks; or previously encoded and reconstructed blocks of the image. The image is encoded in the video encoder using the initial reference blocks and the sets of blocks as predicted.

Abstract: Method and apparatus for intra-prediction in a video encoder are described. An aspect relates to a method of intra-prediction for a group of samples in an image being coded. In some examples, the method includes: defining a target template for the group of samples; comparing the target template with affine transformations of candidate templates within a search area of the image; identifying at least one matching template of the candidate templates as matching the target template; determining a candidate group of samples based on the at least one matching template; and coding the group of samples using the candidate group of samples as a predictor.

Abstract: A method and apparatus for adaptive interpolation filtering for image compression is disclosed. The method includes determining an activity measure associated with a set of pixels neighboring a pixel undergoing intraframe prediction or a distance measure between at least one pixel in the set of pixels and the pixel undergoing intraframe prediction, and selecting a filter for filtering at least a portion of the set of pixels in accordance with the at least one of the activity measure or the distance measure.

Abstract: The color tone compensation method provides a simple and efficient method to compensate the color tone differences between two different sources of images. A first image sample, such as a still image, from a first image capturing source and a second image sample, such as a video frame, from a second image capturing source are aligned, and a tone-mapping estimation routine is applied to the two aligned images. The tone-mapping estimation routine uses the pixel intensity value histograms associated with the two aligned images and generates a tone mapping table. The tone mapping table includes a conversion intensity value for each intensity value in the second image. The conversion intensity value is a statistical measure, such as the mean, calculated according to the data in the corresponding pixel intensity value histogram.