At&t Bell Laboratories

Lucent Technologies Inc.

Other

At+t Corp.

Eric D Petajan

Arun N Netravali

Scott C Knauer

Kim N Matthews

Robert J Safranek

Peter H Westerink

James David Johnston

Hans Peter Graf

Hosagrahar Visvesvaraya Jagadish

Charles Douglas Blewett

Vinod Anupam

George John Kustka

Bruce F Naylor

Kobad Bugwadia

Alireza F Faryar

Scott C Knaur

Narindra N Puri

Shaul Dar

Theodore F Leibfried

A process for up-converting an existing video source signal having a low frequency (frames/second) to a high frequency signal for use with High Definition Television (HDTV). The process samples the existing frames in the existing video signal and calculates integer displacements of pels within the existing frames. A polynomial curve fit is then performed on the displacements to obtain estimates of horizontal and vertical displacements of each block in each existing frame. Based on the alignment of the blocks within a sampling grid on each frame, the blocks are segregated into groups. The block groups are then used to interpolate missing or required frames of the high frequency signal in a piecemeal manner by utilizing blocks of a particular block group to estimate a corresponding block in a frame of the high frequency signal.

Motion compensation image interpolation&mdash;frame rate conversion for HDTV

A face feature analysis which begins by generating multiple face feature candidates, e.g., eyes and nose positions, using an isolated frame face analysis. Then, a nostril tracking window is defined around a nose candidate and tests are applied to the pixels therein based on percentages of skin color area pixels and nostril area pixels to determine whether the nose candidate represents an actual nose. Once actual nostrils are identified, size, separation and contiguity of the actual nostrils is determined by projecting the nostril pixels within the nostril tracking window. A mouth window is defined around the mouth region and mouth detail analysis is then applied to the pixels within the mouth window to identify inner mouth and teeth pixels and therefrom generate an inner mouth contour. The nostril position and inner mouth contour are used to generate a synthetic model head. A direct comparison is made between the inner mouth contour generated and that of a synthetic model head and the synthetic model head is adjusted accordingly. Vector quantization algorithms may be used to develop a codebook of face model parameters to improve processing efficiency. The face feature analysis is suitable regardless of noise, illumination variations, head tilt, scale variations and nostril shape.

Face feature analysis for automatic lipreading and character animation

A method and system for visualizing relatively large amounts of data in a limited display space includes segmenting the data, mapping each segment to a node within the display space based on a specified nodal layout, texture mapping each graphical display of each segment of data to its node, and displaying the data in each node of the display space as mapped. This visualization scheme enables a user to map a relation to a specification of an N-dimensional display, by designating how attributes are to be used to get a desired segmentation, layout and presentation and enables a user to explore the data with various exploring features including real time continuous scaling.

Scalable data segmentation and visualization system

Apparatus and methods are provided for analyzing and enhancing a received data signal representing one or more views of a captured scene. The received data signal includes a plurality of data points. Each view includes a plurality of image points. Ones of the plurality of data points are representative of ones of the plurality of image points. A plurality of data sets are identified wherein each one of the data sets includes ones of the plurality of data points. Ones of the identified data sets are compared with one or more standard values and, in response thereto, are enhanced selectively to generate an output signal representing a processed 3-D scene estimate.

Apparatus and methods for performing electronic scene analysis and

In a hybrid system for synthesizing a compressed video bitstream from a dynamic 3-D model, the video bitstream represents an image sequence comprised of a plurality of image frames which are sequential in time and in which each image frame is partitioned into a set of regions. Motion vectors are generated for each region within a current image frame which represent the dynamic 3-D model. A prediction error is computed for each region based on the generated motion vectors. The type of encoding to be performed on each region in a succeeding image frame is determined based on the value of the prediction error. Each region in the succeeding image frame is encoded as indicated by the determined encoding type.

System and method of generating compressed video graphics images

A method of incorporating an auxiliary signal with an input video signal includes the step of encoding the input video signal to generate a variable bit rate encoded signal having a selected level of distortion. The video signal is applied to a buffer. The level of distortion is selected as a function of the fullness of the buffer. The video signal is read from the buffer and applied to a transmission channel at a predetermined video output rate. Portions of an auxiliary signal are applied to the channel at times when the buffer is in an underflow condition. The selected level of distortion has a flow which corresponds to a predetermined, non-zero level of perceptual distortion in the video signal.

Method of incorporating a variable rate auxiliary data stream with a

An apparatus provides digital data representative of a compressed high definition video image signal in a manner such that a conventional video tape recorder (VTR) employing current electrical and mechanical technologies is adaptable for reading this data for display in special modes such as fast forward, reverse and still frame. To minimize the loss of data while the VTR is operating in one of these special modes, the data representative of the high definition image signal is decoded by the apparatus to yield available time slots in the data stream. These time slots are then filled with data which duplicates the most important of the original transmitted data then existing in other time slots in the data stream. This duplicating of the data increases the likelihood that the important picture information will be read by the VTR while operating in a special mode. To facilitate fast recovery from lost of synchronization, the apparatus utilizes a small fraction of the original image which is included in each section or slice of the transmitted image. As synchronization is reacquired, the high definition video image is built from the image found in the slices.

Apparatus and method for displaying recorded compressed digital high

An improved predictive encoder where the leak signal is a function of the buffer fullness of the encoder. More specifically, the signals stored in the encoder output buffer are further encoded based on the fullness of the buffer, and information about this further encoding is used in determining the leak factor level. In accordance with another improvement, this leak factor level is not constrained to granularity that is imposed by the decoder hardware. Removal of the constraint is accomplished by cycling through a sequence of permissible leak levels that averages at the desired level.

Adaptive leak HDTV encoder

The effects of perturbations in a decoder responsive to a signal encoded in a predictive coding schema are reduced by shortening the time needed to converge the decoder's prediction signal to that of the encoder. This is accomplished by temporarily altering the leak factor when it is known that a perturbation is about to manifest itself and by, in some situations, altering the incoming encoded signal itself.

HDTV receiver

Blocks of HDTV picture information are selected for transmission at a plurality of channel rates of an HDTV transmitter. The size of each block of HDTV picture information is dependent on a target distortion parameter for the HDTV picture information. As a result of this selection, a portion of each block of HDTV picture information is transmitted at the lower channel rate, with the result that the HDTV transmitter range is maximized while maintaining a picture quality for the resulting HDTV video image.

Soft coding for HDTV

A high definition television system that is characterized by low transmission bandwidth is achieved by removing redundancies in the signal, encoding the remaining signals, and transmitting the encoded signal in a manner that is most compatible with the applicable standards. The dynamic range of the encoded signal is reduced by combining groups of adjacent samples that are smaller than a 'floor' threshold to form larger samples, and by splitting samples that are larger than a 'ceiling' threshold to form a plurality of smaller samples. Encoding of the combined samples is carried out in a conforming manner to minimize deleterious effects of transmission noise.

PAM signals modulation with mappings to improve utilization of available

A motion compensated encoder where motion vectors are selected based on the prediction error generated in localized areas of the encoded image and based on an available bit budget. The motion vectors are created by dividing the image into blocks of two sizes and by considering the best mix of large and small size blocks, and their associated motion vectors, that minimize the overall prediction error, within the constraints of the bit budget. For convenience, the image division is arranged so that a given number of small sized blocks forms one large sized block (e.g. 16:1). Also, the block sizes are arranged so that employing only large sized blocks does not exceed the given bit budget, while employing only the small sized blocks does exceed the given bit budget.

HDTV encoder with forward estimation and constant rate motion vectors

A quantizer, with quantization control that is sensitive to input signal characteristics and to output buffer fullness responds to an input signal that is divided into blocks and DCT transformed. The transformed signal is analyzed to develop a brightness correction and to evaluate the texture of the image and the change in texture in the image. Based on these, and in concert with the human visual perception model, perception threshold signals are created for each subband of the transformed signal. Concurrently, scale factors for each subband of the transformed signal are computed, and a measure of variability in the transformed input signal is calculated. A measure of the fullness of the buffer to which the quantizer sends its encoded results is obtained, and that measure is combined with the calculated signal variability to develop a correction signal. The correction signal modifies the perception threshold signals to develop threshold control signals that are applied to the quantizer. The scale factors are also applied to the quantizer, as well as a global target distortion measure.

Adaptive non-linear quantizer

In a differential PCM encoder, the problem of error perpetuation is solved by leaking a changing, rather than a fixed, fraction of the input signal to the differential PCM. The fraction leaked is sensitive to the characteristics of the signal. In one embodiment the fraction leaked is fixed for a frame in accordance with a chosen characteristic of the frame signal. In another embodiment, the fraction leaked is set in accordance with one function when a chosen characteristic of the frame signal exceeds a given level, and follows another function when the chosen characteristic does not exceed the chosen level. In a still another embodiment, the fraction leaked is set to one of two levels, based on a chosen characteristic of the frame signal.

An HDTV receiver design includes a reconstruction section and an inner loop section. The reconstruction section comprises a receiving section for accepting the television receiver's antenna signals, for separating out the component signals from the received signals and for decoding the separated signals. The decoded signals are applied to a quantization decoder that is responsive to vector codebook and to applied quantized vector signals, and the output signals of the quantization decoder are applied to an inverse DCT circuit. The inner loop comprises an adder for adding an estimate signal to the output of the DCT circuit, a frame buffer, a motion compensation circuit that is capable of translating large blocks as well as small blocks and a leak circuit that modulates the output of the motion compensator circuit in accordance with a received leak control signal. The input to the frame buffer is also applied to conventional amplification and display circuits to form the familiar output of a television receiver.

Growing community of inventors

Watchung, NJ, United States of America

Eric D Petajan