Abstract: A method and apparatus for encoding, illustratively, a video information stream to produce an encoded information stream according to a group of frames (GOF) information structure where the GOF structure and, optionally, a bit budget are modified in response to, respectively, information discontinuities and the presence of redundant information in the video information stream (due to, e.g., 3:2 pull-down processing).

Abstract: An apparatus and method for implementing object trajectory segmentation for an image sequence. Specifically, block-based motion vectors for a pair of adjacent frames are used to derive optical flow, e.g., affine, motion parameters. The object trajectory segmenter applies the optical flow motion parameters to form a new prediction or method for predicting the positions of all the points on an object over time within an interval. The new prediction is then applied and the result is compared with an error metric. The results from such comparison with the error metric will dictate the proper intervals (temporal boundaries) of the image sequence at which the motion parameters are valid for various key objects.

Abstract: For video compression processing, each frame in a video sequence is segmented into one or more different regions, where the macroblocks of each region are to be encoded using the same quantizer value, but the quantizer value can vary between regions in a frame. For example, for the videophone or video-conferencing paradigm of one or more “talking heads” in front of a relatively static background, each frame is segmented into a foreground region corresponding to the talking head, a background region corresponding to the static background, and an intervening transition region. An encoding complexity measure is generated for each macroblock of the previous frame using a (e.g., first-order) rate distortion model and the resulting macroblock-level encoding complexities are used to generate an average encoding complexity for each region. These region complexities are then used to select quantizer values for each region in the current frame, e.g.

Abstract: Frames in a video sequence are divided into two or more regions and a specified number of macroblocks are selected in each region for intra-coding. Depending on the particular implementation, for one or more of the regions, the intra-macroblocks are selected randomly, while at least one other region is dividing into a specified number of slices with the least-recently intra-coded macroblock in each slice selected for intra-coding. When an error is detected at the decoder, the decoder discards data in the corresponding packet and applies a concealment strategy that involves using motion-compensated data if the motion vectors were accurately decoded; otherwise, using non-motion-compensated reference data for the macroblocks affected by the discarding of data. The refresh strategy of the present invention can be used to provide the resulting encoded bitstream with resilience to transmission errors, while maintaining an acceptable degree of video compression.

Abstract: Apparatus and method for classifying regions of an image, based on the relative “importance” of the various areas and to adaptively use the importance information to allocate processing resources, e.g., bit allocation in an encoding environment.

Abstract: During video coding, a transform such as a discrete cosine transform (DCT) is applied to blocks of image data (e.g., motion-compensated interframe pixel differences) and the resulting transform coefficients for each block are quantized at a specified quantization level. Notwithstanding the fact that some coefficients are quantized to non-zero values, at least one non-zero quantized coefficient is treated as if it had a value of zero for purposes of further processing (e.g., run-length encoding (RLE) the quantized data). When segmentation analysis is performed to identify two or more different regions of interest in each frame, the number of coefficients that are treated as having a value of zero for RLE is different for different regions of interest (e.g., more coefficients for less-important regions).

Abstract: An image is divided into one or more (e.g., foreground) regions of interest with transition regions defined between each region of interest and the relatively least-important (e.g., background) region. Each region is encoded using a single selected quantization level, where quantizer values can differ between different regions. In general, in order to optimize video quality while still meeting target bit allocations, the quantizer assigned to a region of interest is preferably lower than the quantizer assigned to the corresponding transition region, which is itself preferably lower than the quantizer assigned to the background region. The present invention can be implemented iteratively to adjust the quantizer values as needed to meet the frame's specified bit target. The present invention can also be implemented using a non-iterative scheme that can be more easily implemented in real time.

Abstract: A method and apparatus for adapting and enhancing the behavior of an MPEG-like encoder to the presence and/or absence of 3:2 pull-down processed video information within a video information stream to be encoded. Specifically, a rate controller within an MPEG-like encoder, in response to a detection of 3:2 pull-down processing of material to be encoded, such as mixed mode video material, dynamically allocates a group of pictures (GOP) bit budget.

Abstract: After performing integer-pel motion estimation to select the best integer-pel location, half-pel (or other fractional-pel) motion estimation is performed by estimating the distortion values for all of the surrounding half-pel locations by linearly interpolating using the distortion value for the best integer-pel location and the distortion values available from the integer-pel motion estimation analysis corresponding to the surrounding integer-pel locations. A subset of those half-pel locations is then selected as candidate half-pel locations (e.g., based on lowest estimated distortion values) and true distortion values are then generated for those candidate locations. The best half-pel location is then selected based on the distortion values for the candidate half-pel locations and the best integer-pel location. This best half-pel location may then be used to perform the motion-compensated inter-frame differencing step of a video coding scheme.

Abstract: The algorithm assumes a constant bit rate over a timing window of specified duration (e.g., a specified number of consecutive frames), where the current frame to be encoded lies in the interior of the timing window. A target bit rate for the current frame is initially selected by calculating the number of bits already used to encode other frames within the window and then assuming that the remaining available bits allocated to the timing window will be evenly distributed to the remaining unencoded frames in the timing window. The target bit rate may then be optionally adjusted based on scene content, encoder state, and buffer considerations. Through a combination of target bit allocation and frame skipping, spatial and temporal resolutions are maintained within acceptable ranges while meeting buffer delay constraints. The algorithm has also been extended to include PB frames in addition to P-only coders.

Abstract: An apparatus and concomitant method for selecting a macroblock coding mode based on the measured distortion associated with the selected coding mode is disclosed. The invention selects a coding mode that has a distortion measure that is nearest to an expected operating distortion. Once an initial coding mode is selected, the invention applies a trade-off operation. The best coding mode after the trade-off operation is selected as the coding mode for the current macroblock.