An image motion vector sensor is provided which includes a motion vector sense section in which an input signal of an image plane is subdivided into a plurality of sense domains so as to compute a motion vector of an image in each of the domains, a motion vector decision section for judging validity of the motion vector in each domain so as to decide a motion vector of an entire image plane, a vector reliability density decision section for producing a coefficient which is a function of the density with respect to time of a signal proportional to the number of the domains judged to be valid, and a control section for controlling a magnitude of the motion vector of the entire image plane based on an output from the vector reliability density decision section.

An image signal decoding apparatus for decoding a transmitted image information signal into an original image signal: wherein a one frame image signal constructed of a plurality of image signals is divided into a plurality of blocks each constructed of a plurality of image signals; a block characteristic information signal and a coded information signal are received on a block unit (block-by-block) basis; and in accordance with a decision result by a motion decision circuit, either the image signals in another block corresponding to the same image in a block which was decided as being incapable of being decoded into original image signals, or the image signals of another block corresponding to a different image in the block decided as being incapable of being decided, are selected; and the selected image signals are interpolated as interpolation image signals. Any image motion can therefore be detected at once, and an adaptive interpolation can be effected in accordance with the image motion information.

A movement compensation predictive coding method including: a first step of detecting a movement vector MV.sub.M between the N-th frame (N.gtoreq.0) and the (N+M)-th frame (M.gtoreq.2), the vector MV.sub.M being assumed as an initial value of a difference movement vector MVe, and to detect a movement vector MVi between the (N+1)-th frame (i=0, 1, 2, . . . , M-1) and (N+i+1)-th frame; a second step wherein a detected movement vector MVi exists in the vicinity of a predictive movement vector MVe/(M-i) predicted from the difference movement vector, the predictive vector MVe/(M-i) is assumed as a correct movement vector, and a movement compensation predictive coding is carried out from pixels predicted by using pixels in the vicinity of the coordinate indicated by the predictive vector MVe/(M-i), and obtaining a first subsegment difference movement vector, by subtracting the predictive vector MVe/(M-i) from the difference vector MVe; and a third step wherein when no detected movement vector MVi exists in the vicinity of a predictive movement vector MVe/(M-i), the detected vector MVi is assumed as a movement vector, and a movement compensation predictive coding is carried out using the vector MVi, and obtaining a second subsegment movement vector by subtracting the movement vector from the difference vector MVe.

An apparatus for restoring an original image signal from a transmitted image information signal formed by dividing one picture amount of image signal composed of a plurality of picture element signals into a plurality of blocks each composed of a plurality of picture element signals and high-efficiently coding, for every block, the plurality of picture element signals constituting a respective one of the plurality of divided blocks is arranged to judge whether the picture element signals in each block are restorable or not on the basis of the inputted image information signal, to detect the presence or absence of motion in an image represented by the picture element signals of a block in which the picture element signals have been judged not to be unrestorable, to select either the picture element signals of a block belonging to the same picture as the unrestorable picture element signal block or the picture element signals of a block belonging to another picture according to a result of detection, and to output the selected picture element signals as interpolation picture element signals. The arrangement enables the apparatus to carry out an interpolation process in a manner most opposite to the state of the image to be reproduced.

Motion vectors resulting from block matching are corrected for periodic structures by taking a more reliable vector from an edge of a moving object containing the periodic structure. This is done by calculating and comparing different error combinations to identify periodic structures and replacing a current motion vector with one of an adjacent pixel block either from the block to the left or from the block above, which ever yields the smaller error in the current block, or by taking a combination (e.g., mean) of both vectors. Advantageously, artifacts such as phase reversal of information in interpolated fields due to the presence of an erroneous vector in a periodic structure are reduced and very little additional processing is required after the block matching itself.