| 摘要 |
<p>PURPOSE:To improve detection accuracy in respective hierarchies by detecting the maximum value of picture element values provided in the block of a low-order hierarchy for the respective blocks of the low-order hierarchy and preparing a high-order hierarchy at the time of successively detecting a motion vector in the low-order hierarchy with high resolution based on the motion vector detected in the high-order hierarchy with low resolution. CONSTITUTION:In this moving amount detector 1, by detecting the maximum value and a minimum value from the prescribed blocks on an original picture which is the lowest hierarchy, the high-order hierarchy is generated. For instance, by obtaining the maximum value inside the block composed of (two picture elements)X(two lines) of the lowest hierarchy, one picture element of the hierarchy one order higher is prepared. Thus, four maximum values of a second hierarchy are obtained from 16 picture elements of a first hierarchy and further, the maximum value is obtained from 4 picture elements of the second hierarchy in the hierarchy higher than the second hierarchy, that is a third hierarchy and the value is defined as one picture element of the third hierarchy. Similarly, by obtaining the minimum value inside the block composed of the (two picture elements)X(two lines) of the lowest hierarchy, one picture element of the hierarchy one order higher, that is the second hierarchy, is prepared and one picture element of the third hierarhcy is defined.</p> |
