Abstract: An image recognition system uses feature or attribute matching for fast coarse selection of a group of reference images which closely resemble an unknown image and thereafter selects the best reference image match from that group using 2-dimensional intensity correlation or template matching. The system provides highly accurate image recognition without resorting to time consuming exhaustive search by template matching.

Abstract: In a transmitter for image signals, a variable-length coder whose code set has been optimized for the order statistics of a run-length-coded bilevel image is used to code a stream of differential-pulse-code-modulated (DPCM) errors representing a multi-level image. The DPCM errors are first rearranged on a line-by-line basis in such a way that the rearranged error stream contains many long runs. Individual words in the rearranged error stream are then changed in such a way that, for at least one value of n, the n most-frequently-occurring words in the resulting, processed signal substantially have the aforementioned order statistics. The resulting signal is then run-length-coded, applied to the variable-length coder, and then transmitted. The reverse processing is performed at the receiver.

Abstract: This disclosure relates to the instant method and system for controlling the interconnecting of a plurality of local communications networks. A source station and a destination station may respectively be adapted to be coupled to at least one communications channel of a communications path, which path includes a plurality of channels, each channel being assignable to at least one of a plurality of local data networks. The respective stations are adapted to determine the home network of the other respective stations with which the respective stations may communicate. Broadly, a source station, which has a message to transmit to a destination station, determines the home network of the destination and couples a first source receiver and a source transmitter to the home network of the destination for transmitting a packet from the source to the destination. A first packet may be transmitted when the source station acquires access to an idle channel of the destination home network.

Abstract: The instant communications system and method for controlling both digital voice traffic and digital data traffic includes a communications path, a source station and a destination station. The source and destination stations are respectively adapted to be coupled to a source station home network and to a destination station home network. The source station is further adapted to transmit a signaling packet on a first transmit control channel. The signaling packet is for signaling a destination station that the source station has an information packet intended for the destination station. The destination station is adapted to detect the signaling packet on a second receive control channel. The signaling packet includes a field for identifying the source station which has the information packet intended for the destination station. In response to the signaling packet, the destination station is coupled to the source home network.

Abstract: When a DPCM encoded video signal is spatially subsampled in order to reduce the amount of information that must be transmitted, the values of nontransmitted picture elements (pels) must nevertheless be reconstructed at both transmitter and receiver. In accordance with the present invention, reconstructed values are obtained by interpolation between nearby transmitted elements only in unpredictable ones of the picture. In regions where prediction errors are low, the predicted value of a nontransmitted element is instead used as a basis for the reconstruction. Increased efficiency is thus achieved at little increase in complexity, since the same prediction required for DPCM encoding is used for reconstruction of nontransmitted pels.

Abstract: A ventricular tachycardia (VT) prediction system for use with an ECG input having cyclical characteristics. The ECG input is periodically sampled during each of a plurality of ECG cycles, and signals representative of the ECG input for each corresponding sample time are stored. A high pass filter adaptively filters the signals for each ECG cycle, so as to provide a filter output systematically related to the high frequency content of the ECG input. The adaptive filter produces segments which are uninfluenced by other samples so as to negate ringing. A reference signal is provided for VT prediction. The filter output is compared to the reference signal, and the prediction of ventricular tachycardia is indicated responsive to the comparison. In one embodiment the samples are averaged for a plurality of ECG cycles prior to generating the filter output. In another embodiment, the system determines the duration of the QRS portion of the ECG cycles prior to generating the filter output.

Abstract: The flicker which results from large intensity differences in adjacent scan lines of a symbol displayed in an interlaced-field format is reduced by a non-linear signal filter and signal generator. The disclosed filter and signal generator changes the intensity of a scan line adjacent to a scan line of a symbol in response to a detected predetermined intensity difference between the adjacent scan line and the symbol scan line.

Abstract: DPCM or PCM samples which have been quantized at a transmitter are assigned a representative value associated with one of a plurality of mutually exclusive quantization levels, each having predefined upper and lower limits. The sample values are reconstructed at the receiver using (in addition to the representative value) information regarding the quantizer characteristics as well as information derived from spatially or temporally correlated samples. In DPCM systems, the same technique can also be used at the transmitter to improve prediction of the present sample by improving reconstruction of previously processed samples.

Abstract: A method and apparatus for the transmission of two-tone pictures progressively in several stages. Progressive transmission enables a viewer to receive and utilize pictures in a small fraction of the normal transmission time. In a first stage, information representing groups of picture elements of preselected lines of the scanned picture is transmitted which allows a coarse version of the picture to be reproduced at a receiver. During the subsequent stages the horizontal resolution of the previously transmitted scan lines is improved and additional scan lines are transmitted in a specific order until the received picture is an identical copy of the original picture. The progressive picture information is efficiently transmitted in one of several line-by-line codes resulting in a total transmission time not significantly longer than the time for a nonprogressive transmission.

Abstract: A pseudo-gray scale signal obtained by comparing the intensity values of elements of a picture with a series of spatially dependent threshold values is processed for display on a multilevel device. Generally, an estimate of the intensity value of each pel is formed as a joint function of the display words which make up the pseudo-gray scale signal and an indication of the threshold value associated with the pel. In one embodiment, the display words and threshold values are used to select appropriate stored intensity values obtained from representative pictures. Alternatively, the estimates are derived recursively using display words and threshold values for a neighborhood near the pel being processed.

Abstract: Information defining elements of a picture is estimated by interpolation using information from related locations in preceding and succeeding versions of the picture. The related locations are determined by forming an estimate of the displacement of objects in the picture. Displacement estimates are advantageously formed recursively, with updates being formed only in moving areas of the picture. If desired, an adaptive technique can be used to permit motion compensated interpolation or fixed position interpolation, depending upon which produces better results.

Abstract: The efficiency with which a video signal is encoded is increased by examining successive blocks of M picture elements to determine which K pels have frame differences which do not exceed a predetermined threshold. The values of these K pels are intentionally altered, and the pels are applied to a transform circuit which generates N.ltoreq.M corresponding coefficient values. Only N-K of the coefficient values, and a code indicating the unchanged pel locations are then encoded for transmission. The decoder (and the feedback loop of the encoder) is arranged to recover the original picture using only the frame differences for pels which have changed significantly.

Abstract: Pictorial information is encoded using an interframe prediction technique which accounts for spatial and/or temporal changes in illumination. In one embodiment (FIG. 3), the intensity value of each element of a picture is predicted using the intensity of the corresponding element in a previous frame and a recursively updated gain factor .rho. which can account for both illumination changes as well as object movement. The prediction error and pel location is encoded only if the error exceeds a predetermined threshold. In a second embodiment (FIG. 3), additional predictions are made using displacement compensated prediction alone and in combination with gain adjustment. The best predictor is selected and the corresponding prediction error transmitted only if it is large.

Abstract: Video signals are encoded (FIG. 5) using motion compensated prediction which operates on a transform domain representation of the signal. The displacement estimate used for compensation is recursively updated, so that a feedback path exists between the output (205) of the displacement calculator and one of its output (209). The update term is also computed in the transform domain. A decoder (FIG. 6) uses the same prediction technique to recover the original picture.

Abstract: Supplementary digital data is encoded together with a video signal by predicting the intensity values of a first set of video samples, compressing the range of the prediction errors and superimposing a value indicative of the supplementary data on the compressed value. The samples in the first set may be selected in a regular sequence or asynchronously in accordance with a function of the video signal such as the slope in the area of the picture being processed; the composite values are time multiplexed with the remaining intensity values.

Abstract: Method and apparatus for encoding a video signal wherein the prediction errors for previously encoded pels are evaluated using motion compensation and frame difference predictions. Each present pel intensity value is then predicted using the technique which produced the smaller prediction error sum. Prediction errors for the present pel are encoded, together with address information, if the error exceeds a threshold value. The motion compensated prediction utilizes a displacement estimate which is recursively updated. A decoder is also described.

Abstract: A technique for estimation of displacement and/or velocity of objects in television scenes is recursive, in that each displacement calculation is made using the results of the previous calculation. In a first embodiment, a displaced frame difference DFD (i.e., the intensity difference between a spatial location in the present video frame and a location in the previous frame displaced from that location by an initial displacement estimate) is updated using the initial value of DFD normalized by a measure of local picture activity. In a second embodiment, the update is a multiplicative function of DFD and the intensity gradient at the displaced location in the previous frame.

Abstract: A motion compensated video encoder includes apparatus (100) for recursively updating an estimate of displacement in the moving area of a picture, and for predicting the intensity values of elements of the picture using the displacement estimate. Address information is encoded for background (non-moving) elements and for moving elements for which the motion compensated prediction errors are satisfactorily small. For uncompensable pels, where the error exceeds a threshold, both address information and error values are encoded. The displacement estimate can also be encoded, if desired.

Abstract: A series of bits which represent the intensity values of elements of a facsimile picture are preprocessed to eliminate runs of length one, and then applied to a predictive encoder which treats values for alternate pels in two distinct ways. Specifically, the value of every odd pel is predicted, based upon the intensity values of some surrounding picture elements. Each prediction is compared with the true intensity value to yield an error signal. The encoder is also arranged to generate a reference signal which indicates whether the confidence in each prediction for odd pels is high or low. The error values for odd pels within a predefined reordering interval are then sorted or reordered into two groups, dependent upon the reference value. On the other hand, the intensity values of even pels are predicted only when the values cannot be precisely determined from the two previous and subsequent values; the error values for even pels are assembled into a third group.

Abstract: A facsimile signal encoding technique reduces the amount of information needed to represent a picture by (a) generating an error signal (on line 1120) which indicates whether an intensity value prediction corresponds to the true intensity value, (b) generating a reference signal (on line 1005) indicating whether the confidence in each prediction is high or low, (c) reordering the error signal values (logic 1010) into high and low confidence groups in accordance with the reference signal, and (d) assigning code words (encoder 1016) to represent the run lengths of the reordered data. Advantageously, different code dictionaries are selected (line 1019) for code assignment in each of the groups, each dictionary being tailored to the type of runs expected in each group. Also, code words for certain runs may be dropped, and the lengths of other runs may be combined (by logic 1015) before run length coding, again increasing encoder efficiency. A receiver (FIGS. 16-18) for decoding the run lengths is described.