| 摘要 |
In a video coding/decoding system, a high-resolution reference image having pixels in closer relation is produced by applying the interpolating operation of a luminance value to a reference image. In the matching process for motion estimation, when the luminance value of a point lacking a pixel in the reference image becomes required, the luminance value of a pixel at a point nearest to the high-resolution reference image is approximated without computing the interpolation. As a result, the number of times the interpolation of the luminance value is computed can be reduced, thereby reducing the circuit scale or the processing time. Also, the horizontal and vertical components of the motion vector of each pixel used for synthesizing a predicted image are defined to assume only a value equal to an integer multiple of 1/d1 and 1/d2 (d1 and d2 being position integers) of the distance between adjacent pixels. Consequently, a mismatch of a predicted image attributable to the difference of the computation accuracy of the motion vector in a simple computation is prevented thereby to simplify the process.
|
