Abstract: A method and system for compression encoding a video signal representative of a plurality of image frames which include a starting frame, an ending frame, and at least one intermediate frame, provides for identifying key points of at least one object in the starting and ending frames. For each object, a mathematical transformation is determined which estimates a representation of the corresponding key points in each intermediate frame by mapping the feature points from the starting frame to the ending frame. The representation of the object in the starting frame and the mathematical transformation are encoded and multiplexed to form an object bit stream. Each object bit stream produced in response to a set of image frames is then multiplexed to form an image bit stream. A signal-to-noise ratio can be scaled by adjusting the number of feature points identified, which in turn allows the compression ratio to be selectively changed.

Abstract: A system and method for encoding the speech of multiple speakers for transmission on a limited medium such as ISDN or ADSL. The system and method encode one or more “dominant” speakers on a first channel, and merges all other speakers into a second channel. A “dominant” speaker may be defined according to the needs of the application, but intuitively it is a speaker in a collaboration session who has the “floor”. The method and system transmit speech through a teleconferencing system having an available bandwidth. Speech from a speaking one, or ones, of the plurality of participants is passed while speech from non-speaking participants is rejected. The system responds to the speech from the one, or ones of the speaking participants passed by the speech detector, and determines which of the detected speaking one, or ones, thereof is a dominant speaker.

Abstract: Methods and systems for encoding a video signal representative of a plurality of image frames which include a starting frame and at least one subsequent frame, include a step of identifying at least one object in the image frames. A plurality of feature points of the at least one object are located in the image frames. A mathematical transformation which maps locations of the feature points in the starting frame to locations in each subsequent frame is determined. The bit plane representation of the starting frame and the mathematical transformation are encoded and multiplexed to form a bit stream. Methods and systems for decoding the bit stream perform a step of transforming the bit plane representation of the starting frame in accordance with the mathematical transformation to form a bit plane representation for each subsequent frame.

Abstract: A system and method for communication of information using channels of different latency combine a high latency communication channel with a low latency communication channel to reduce the communication delay perceived by a user. The system and method includes separating information into first and second components based on a parameter correlated to the perceived delay, communicating the first component via a first channel, communicating the second component via a second channel having a communication delay greater than the first channel, and generating a representation of the information based on the first component. The second component of the information may be used to augment or modify the information represented by the first component.

Abstract: A method is provided for reducing the number of data bits required for storage or transmission of the data representing the alpha masks of objects in a single image or frame of a video sequence by jointly encoding overlapping boundaries between two adjacent objects within the image and by avoiding redundant encoding and transmission of an alpha mask which can be reconstructed from other encoded alpha mask data.

Abstract: Methods and systems for compressing data representing a video sequence having a plurality of image frames include obtaining a plurality of basis image frames which represent an anticipated range of motion of at least one object in the video sequence and an actual current image frame and a previous image frame of the at least one object. A plurality of feature points of the at least one object are located in each of the plurality of basis image frames, the previous image frame, and the actual current image frame. A mathematical transformation which maps locations of the feature points in each of the plurality of basis image frames and the previous image frame to locations in the actual current image frame is determined. The mathematical transformation is then encoded for transmission.

Abstract: A method for residual coding of model failure regions in compressed video sequences in which a database of residuals for previous model failures is maintained at both the encoder and the decoder. The residuals for each determined model failure regions are compared with the residuals for previous model failures for the same region. In the event a current residual is identified in the database, only the index to the residual in the database needs to be sent. An additional residual taken after the model failure region has been partially corrected with a current residual identified in the database may also be transmitted. This mode provides increased quality at reduced bit rates by maximizing the utilization of previously transmitted or stored data. The process continuously updates the database at both the encoder and decoder when the current residual is not found. The process uses residuals of residuals to provide increased quality of reduced bit rates.

Abstract: A method for estimating high frequency components of a compressed image normally lost during a data compression stage and an encoder and a decoder for carrying out same. The image is compressed to obtain unquantized and quantized frequency domain coefficients. An inverse quantization is performed on the quantized coefficients to obtain lossy frequency domain coefficients. Loss characteristics of the high frequency components are then determined based on the unquantized frequency domain coefficients and the lossy frequency domain coefficients according to one of several methods. Finally, the loss characteristics are coded to obtain encoded loss data corresponding to the lost high frequency components.