Abstract: In a multichannel encoding method for encoding digitized pictures, in particular digitized stereoscopic television pictures, the digitized pictures are processed in different channels comprising at least a movement estimation channel to provide digitized picture encoding data. A particular processing channel is chosen on the basis of this encoding data. In the movement estimation channel an input digitized picture is divided into blocks of pixels to provide a movement vector for each block and an error for each block pixel. For each block, a set of values is calculated relative to a block mean error, a block error standard deviation, a block error mean deviation and an amplitude of the block movement vector to choose the particular processing channel assigned to said block on the basis of the set of values.

Abstract: Disclosed is a method to achieve a sub-sampling of pictures, making it possible to optimize the spatial and temporal definition of the compressed signal, so as to enable a most faithful possible reconstruction of the picture received with said compressed picture signal and, if necessary, with complementary assistance data. A field of speed vectors is defined for at least each picture to be sub-sampled, and said pictures to be sub-sampled are sub-sampled by eliminating certain picture points, the preserved picture points being chosen so as to avoid selecting a picture point which corresponds spatially to the approximate projection of a source point preserved during the sub-sampling of the preceding sub-sampling picture, said projection resulting from the application, to said source point, of the associated shift vector.

Abstract: The stereoscopic convergence is adjusted in stereoscopic imaging equipment to increase or decrease the stereoscopic convergence angle. Two imaging devices are disposed symmetrically relative to a median optical axis to generate video signals representing images of an observed scene. Depending on the required stereoscopic convergence angle, two image "read windows" are positioned symmetrically relative to the median optical axis, each being defined by a first pixel and a last pixel of a processed image line. The first and last pixels of a processed line are disposed between the first and last pixels of a raw (oversize) image line, with the result that the video signals generated by the stereoscopic imaging equipment represent righthand and lefthand processed images in corresponding relationship to the raw image read windows.

Abstract: A three-dimensional television system uses a method of processing and transmitting a sequence of pairs of stereoscopic television images by means of a transmission link comprising and analog video channel associated with a digital channel to reconstitute at the receiving end a compatible image or a three-dimensional image. The analog video channel transmits a compatible video image formed by multiplexing image items of righthand and lefthand images of pairs of stereoscopic images from the sequence.

Abstract: A method for the estimation of the motion of electronic picture points, in the plane of the picture, by means of fast, convergent and adaptive processing, notably in order to enable subsequent, efficient processing, by sub-sampling, of the picture sequence without harming the psycho-visual quality of the picture reconstructed from the sub-sampled picture. To each electronic picture point, there is assigned a displacement vector, computed according to a WALKER and RAO type of algorithmic process of prediction/correction, the predicted value of motion of the current point being the value of the original displacement vector associated with a point of origin of the preceding picture, so that said current point is the approximate projection of said point of origin along said original displacement vector.

Abstract: With interlaced field scanned TV, a point (P) in a field at time (t+1) is determined on the basis of points situated in said field on the same vertical (C and D), and from points situated in the fields which are adjacent in time (t and t+2). The invention provides for multi-gradient motion detection, i.e. relating not only to the points on the same vertical (C and D), and to the points (A and B) having the same position as the point (P) but in the adjacent fields, but also on the basis of points (E, F, G, and H) which are vertically adjacent to the points (A and B) in the adjacent fields and occupying the same fields. Pairs of such points on either side of the point (P) can thus be defined. Interpolation is performed along each of the available pairs, and the interpolations are weighted as a function of the corresponding gradient.

Abstract: With interlaced field scanned TV, a point (P) in a field at time (t+1) is determined on the basis of points situated in said field on the same vertical (C and D), and from points situated in the fields which are adjacent in time (t and t+2). The invention provides for multi-gradient motion detection, i.e. relating not only to the points on the same vertical (C and D), and to the points (A and B) having the same position as the point (P) but in the adjacent fields, but also on the basis of points (E, F, G, and H) which are vertically adjacent to the points (A and B) in the adjacent fields and occupying the same fields. Pairs of such points on either side of the point (P) can thus be defined. Interpolation is performed along each of the available pairs, and the interpolations are weighted as a function of the corresponding gradient.