Abstract: A half-pixel processing apparatus of a macroblock for motion compensation of a moving image in which NxN pixel data selected for the motion compensation of the moving image is divided into a predetermined unit and half-pixel processed. The half-pixel processing apparatus includes a first unit which sums successive pixel data constituting the N kinds of pixel data and pixel data horizontally adjacent to the successive pixel data constituting the N kinds of pixel data according to a horizontal half-pixel processing signal which is input from an external source. A pipeline unit sequentially stores and outputs the data applied from the first unit. A second unit sums the pixel data output from the pipeline unit and pixel data vertically adjacent to the pixel data output from the pipeline unit according to a vertical half-pixel processing signal which is input from an external source.

Abstract: Combination head-protective helmet which includes a cap and an outwardly extending brim and thermal imaging apparatus which includes an infrared camera for producing an infrared image of a scene or object and a display system which generates a visible image of the scene or object from the infrared image, the camera is mounted underneath the brim of the protective helmet for protection against e.g. falling objects and the display system is mounted on the helmet in a position to permit a person wearing the combination to see the visible image.

Abstract: A method and apparatus for coding and decoding an image signal, and also a recording medium for storing a coded image signal wherein low resolution image signal including a plurality of pixels and also a high resolution image signal including a plurality of pixels are received. The low resolution image signal is coded thereby generating a resultant coded low resolution image signal. The high resolution image signal is predicted from the low resolution image signal thereby generating a predicted high resolution image signal. After that, the predicted high resolution image signal is divided into a plurality of blocks each including n.times.m pixels. The feature of each block is detected and a judgement code corresponding to the detected feature is generated. Then a coding table is selected in accordance with the judgement code. The n.times.m pixels in the corresponding block of the high resolution image signal are coded using the selected coding table to generate coded data.

Abstract: A video encoder for quantizing format-transformed source input-pictures into quantized data, and for coding the quantized data into coded bitstream, for outputting the coded bitstream, comprises a plural-density format-transformer for transforming a source input-picture into a specific format-transformed source input-picture; and a detector for detecting a variation in one of a specific luminance and color difference signal, and for selecting the specific format-transformed source input-picture for input to the quantizer depending on the variation. A decoder having a corresponding configuration is also provided.

Abstract: An image signal interpolating apparatus with a simple structure which generates an interpolation picture element value near the real value without aliasing noise is realized. Predictive coefficients D8 which correspond to the class are read-out, and interpolation data D2 are obtained based on the predictive coefficients D8, thereby, interpolation data D2 near the real value are obtained without aliasing noise. In addition, the flatness near an interpolation-addressed picture element is detected and picture elements used for classification are selected depending on the resultant detected flatness D4, thereby, the interpolation-addressed picture element is classified correctly with the least number of class. Thus the structure can be simplified.

Abstract: An H. 261 encoder outputs coded video data together with a mode selection signal and a judging signal as to discarding or not of a packet. A packetizer receives the output signal from the H. 261 encoder and separately packetizes motion data and DCT data in the input coded data. A gateway judges whether packet including the DCT data is discarded or not with reference to the packet discarding judging signal and also to the bitrate of the network to which the gateway is connected. The gateway thus can effect bitrate conversion according to the bitrate of the network with a simple construction. The gateway may selectively discard high frequency component packets without discarding low frequency component packets, or it may not discard first importance component packets having greater influence on picture quality and selectively discard second importance component packets having less influence.

Abstract: A temporal compression and decompression system is disclosed for color video. A video stream with red, green and blue pixel element values is input to a first computer. In the first computer, a new frame of video data is compared to the frame being displayed on a second computer and the similarity between corresponding blocks and sub-blocks is represented with one tolerance result per sub-block. Based on the number of sub-blocks for which tolerance results exceed a preset threshold and a user-specified quality level, an update quality level is selected that determines the quality at which the frame update is transmitted from the first computer to the second computer. Before the frame update is transmitted the updating information is compressed in a two-stage process. In the first stage 5-bit red, blue and green element values in sub-blocks being updated are is converted to 3-bit categories. In the second stage groups of categories are converted to variable-length Huffman words.

Abstract: A method of motion compensated frame interpolation wherein an intermediate image frame is generated between two image frames in a motion image sequence. The method includes the steps of: identifying a moving foreground or object and the background for two successive frames in a motion image sequence; estimating a velocity vector field for the foreground, and a velocity vector field for the background; using the estimated velocity vector field for the foreground to identify the location of the foreground and the background in the intermediate image; and obtaining the intermediate image by using the foreground velocity vector field to perform motion compensated interpolation for the foreground in the intermediate image and using the background velocity vector field to perform motion compensated interpolation for the background in the intermediate image.

Abstract: An apparatus for statistically multiplexing a plurality of video signals is provided with a complexity detector, coupled to receive a pre-compressed video signal, for generating a complexity signal which relates to the complexity of the pre-compressed video signal, an encoder for compressing a non-compressed video signal at a variable compression rate to generate a compressed video signal, a controller for controlling the variable compression rate of the encoder based on the complexity signal, and a multiplexer coupled to receive the pre-compressed video signal and the compressed video signal. The apparatus may be provided in a broadcasting system, such as a satellite broadcasting system. The broadcast system may include a transmitter coupled to receive the video signals from the multiplexer, a satellite coupled to receive the video signals from the transmitter, a receiver for receiving video signals from the satellite, and one or more decoders connected to receive video signals from the receiver.

Abstract: A method and an apparatus for tracking a target wherein a shape of a target is extracted from a former image which includes an initial image of the target for tracking, and then an area including the shape of the target is extracted as a correlation area. Based on the correlation area, a checking area is extracted from a current image to check for a correlation. By utilizing a correlation function, the correlation between the former image and the current image is calculated. A location change of the target is calculated to estimate a movement of the target, on the basis of the calculated correlation. Compensating the estimated movement during operation of a camera, the camera is controlled so that a gazing point of the camera can track the target.

Abstract: The stereoscopic CG image generating apparatus and a stereoscopic TV apparatus, has a projection transformation section which, based on three-dimensional structural information describing a three-dimensional shape of an object, generates a plurality of two-dimensional projection models as viewed from a plurality of viewpoints, a distance information extraction section which generates a camera-to-object distance information used for calculations in the projection transformation section, and a camera parameter determining section which, based on the output of the distance information extraction section, the screen size of a stereoscopic image display device for displaying finally generated two-dimensional projection models, and a viewer's viewing distance, determines camera parameters so that stereoscopic CG images will be brought within the viewer's binocular fusional range.

Abstract: An MPEG decoder system and method for decoding frames of a video sequence. The MPEG decoder includes motion compensation logic which analyzes motion vectors in an encoded frame of the MPEG stream and uses prior decoded reference blocks to recreate the data encoded by the motion vector. The MPEG decoder stores reference block data according to a novel skewed tile arrangement to minimize the maximum number of page crossings required in retrieving this data from the memory.

Abstract: The optimum predictive image is selected as follows for the entire area of an image sequence comprising a luminance signal expressing each pixel value, and a transmittance signal expressing the transparency of each pixel value, and described by the transmittance signal of the target image to be coded as the area to be coded. The target image to be coded is predicted from at least one template typical of the image sequence, thereby generating a first predictive image. The target image to be coded is also predicted from the image displayed chronologically before the target image, thereby generating a second predictive image. It is then determined whether the first or the second predictive image has the least difference to the target image to be coded, and the predictive image with the least difference is selected as the optimum predictive image.

Abstract: A detection system for detecting a target flying over water, which includes a detector positioned to receive radiation reflected by the target off water, and producing a detector output signal having an amplitude proportional to the amount of received radiation, a high-pass filter for passing portions of the detector output signal having a frequency greater than a prescribed threshold frequency, and for blocking portions of the detector output signal having a frequency lower than the prescribed threshold frequency. The prescribed threshold frequency is sufficiently high to ensure that only the portions of the detector output signal produced by the radiation reflected by the target off the water are passed by the high-pass filter, and that portions of the detector output signal produced by solar glitter are not passed. In a specific embodiment, the system is designed to be installed on a ship for the purpose of detecting low-flying, fast-moving targets, such as cruise missiles.

Abstract: The running real time length of combined video and audio signal programs is shortened or lengthened by deleting or repeating individual fields or frames and corresponding amounts of audio segments. The video and audio portions of the programming material are separated and subjected to processing through a pair of program time changing units. The video portion is processed by deleting individual fields or frames on a manual, periodic or automatic basis. Manual deletion is done by an operator observing the program material on a monitor. Periodic deletion is performed automatically after the operator specifies the total amount of time (or number of fields or frames) to be deleted, with every ith frame or field deleted regardless of content. Automatic deletion is done in a fashion similar to periodic deletion, but the fields or frames are examined and are deleted on the basis of the amount of between frame motion.

Abstract: A video signal is processed to selectively permit copying by superposing in that portion of the video signal which does not contain useful picture information a copyright information signal indicative of whether the viewable picture that is displayed from the video signal is subject to copyright and a copy generation signal indicative of the number of successive generations of copies that can be made from the video signal.

Abstract: (M-N) 0 data are inserted for every N data of the f.sub.SL input data by an interpolation circuit 20 for generating f.sub.SH rate data, while filter coefficients are sequentially generated by M coefficient generators 30A to 30D at the f.sub.SH rate. A register postfix type transversal filter 40 effectuates up rate conversion of N:M (N<M) of generating f.sub.SH rate output data from f.sub.SL rate input data. The transversal filter 40 includes M multipliers 41A to 41D for multiplying the f.sub.SH rate data generated by the interpolation circuit 20 with the filter coefficients sequentially applied by the coefficient generators 30A to 30D and each (M-1) delay circuits 42A to 42C and additive units 43A to 43C for delaying product outputs of the multipliers 41A to 41D by unit time delay and summing the delayed product output together.

Abstract: For image transmission, image data is compressed block by block, and the compressed data is transmitted and decompressed. Due to an information loss during transmission, discontinuities are visible in pictures thus restored, particularly at block boundaries ("blocking effect"). The higher the bit rate, the more visible the discontinuities will be. This is improved by refining the prior art filter. Information for reducing the error is derived from the boundaries of the block, and the filter size is enlarged to 2.times.n+1 pixels, with n=2, 3, 4 preferably limited to one half the number of pixels on one side making up the block. Besides the signal and noise activity of a block, the filter determines both the pixels inside the block and the pixels outside the block. The determination is made by means of the one-dimensional filter in the horizontal and vertical directions.

Abstract: An automatic tracking device is disclosed that extracts data from the comite video output signal of a camera and uses it to electronically isolate an object in the image information on the video output signal and follows the object with a crosshair added to the video by the autotracking device. The autotracking device generates steering signals that are applied to the camera directing circuits to cause the camera to automatically follow the object as the object moves.

Abstract: A method, for use in a block based video signal coding system, for determining a search grid by dividing a video object plane of an object into a plurality of equal-sized search blocks, generates a plurality of search grids, each of the search grids being relatively shifted from its adjacent search gird by a predetermined spacing, and detects edge blocks included in each search grid. Then, the method selects a search grid based on the number of the edge blocks and the pixel values in the edge blocks.