Abstract: A video server for transmitting video across a network to a plurality of users includes a coder (112) which operates in response to call requests or instructions received across the network (102) from individual clients (104, 106). The server maintains in memory a series of user status records (116), one for each of the users. Each record includes, or refers to, a copy of a previous frame sent to the respective user. When the coder (112) receives a call request, it uses the previously-sent frame for that user to optimise coding the current frame that is to be sent, received from a frame grabber (110). The invention avoids the difficulties of video broadcasting while still requiring only a single instance of a video coder.

Abstract: A method and apparatus for data compression comprises applying a decorrelating transform to multi-dimensional data to be compressed, then using a sequence of one or more one-dimensional matching pursuits algorithms to code the output of the transform. The invention finds particular application in video and still image coders, particularly real-time coders, those intended for use at very low bit rates, and those for which scalable bit rates are of importance.

Abstract: A method of data compression, for example for image or audio signals, comprises encoding the data using a transform such as a wavelet transform or DCT, and transmitting only a predefined number of the most significant bits for each coefficient. It has been found experimentally that while such a method may marginally decrease the signal to noise ratio, the subjective quality of the received image or audio signal is nevertheless enhanced.

Abstract: A method of data compression, for example for images or audio signals, comprises encoding the data using a transform such as a wavelet transform or a DCT, grouping the transform coefficients by most significant bit, and/or bit value, and transmitting them group by group. Where the transform uses a code book, such as matching pursuits, the coefficients may further be grouped by code book value. In one embodiment, the method allows the data to be described entirely by the positions of the coefficients, and the groupings, without any need to store or transmit any coefficient value information whatsoever.

Abstract: A method of video motion estimation is described for determining the dominant motion in a video image. The dominant motion is defined by a parametric transform, for example a similarity transform. In the preferred embodiment, selected pairs of blocks in one frame are traced by a block matching algorithm into a subsequent frame, and their change in position determined. From that information, an individual parameter estimate is determined. The process is repeated for many pairs of blocks, to create a large number of parameter estimates. These estimates are then sorted into an ordered list, the list is preferably differentiated, and the best global value for the parameter is determined from the differentiated list. One approach is to take the minimum value of the differentiated list, selected from the longest run of values which fall below a threshold value. Alternatively, the ordered list may be examined for flat areas, without explicit differentiation.

Abstract: A method of video motion estimation is described for determining the dominant motion in a video image. The dominant motion is defined by a parametric transform, for example a similarity transform. In the preferred embodiment, selected pairs of blocks in one frame are traced by a block matching algorithm into a subsequent frame, and their change in position determined. From that information, an individual parameter estimate is determined. The process is repeated for many pairs of blocks, to create a large number of parameter estimates. These estimates are then sorted into an ordered list, the list is preferably differentiated, and the best global value for the parameter is determined from the differentiated list. One approach is to take the minimum value of the differentiated list, selected from the longest run of values which fall below a threshold value. Alternatively, the ordered list may be examined for flat areas, without explicit differentiation.

Abstract: A method of image compression includes significance switching of DCT coefficients in block-based embedded DCT procedures. Bitwise digitized DCT coefficients are passed through successive significance sweeps of the whole image from the most significant down to the least significant coefficient bit planes. With each new sweep, newly significant coefficients may appear within a block, and block-masking is used to transmit the addresses of those newly significant coefficients. An off-mask may also be used. The invention further relates to a hardware or software-based image encoder.