| 摘要 |
<p>The signal converting apparatus and method predictively produces highly accurate interpolated pixels in accordance with a classification which precisely reflects a variety of signal characteristics of inputted video signals to provide a high resolution video signal. An activity is evaluated and classified (12) for each block of an inputted video signal (S1). Stepwise classifications (p0-p2), (Fig 5, 22-24) are executed on each block of the inputted video signal (S1) and a classification (C1) is selected in accordance with an activity code (c0) obtained as a result of the activity classification (21). In this way, the accuracy of subsequent classifications can be increased, reflecting the activity characteristic of each block of the inputted video signal (S1), thus achieving, as a whole, a highly accurate classification of the inputted video signal (S1). Appropriate prediction coefficients (d1) stored in a ROM (14) are read out based on the activity code (c0) and the class code (c1) for each block of the inputted video signal (S1). The coefficients are used in calculators (13) to produce highly accurate interpolated pixel values (d2-d5) which are selected by a selector 15 to provide a high resolution video signal (S2). <IMAGE> <IMAGE></p> |
