| 摘要 |
PURPOSE:To efficiently encode contour information by partitioning an extracted contour line into several blocks and repeating processing until an area extraction evaluated value in the case of regarding a curved line approximated to the contour line composed of those blocks as the contour line is smaller than a threshold value. CONSTITUTION:The image signal of an input image is inputted to a contour extraction processing part 22, and a contour shape is calculated. The contour shape is inputted to an area dividing part 23, and and area shape is calculated by using an evaluation function and divided by a block dividing part 24 while using any suitable reference. Based on the divided block contour shape, an approximate curved line is calculated by a block approximating part 25. The provided approximate curved line is collated with an area division evaluation function and when the evaluated value is larger than the threshold value, it is inputted to the block dividing part 24. Then, the block contour shape is divided again and processing is repeated. When the evaluated value is smaller than the threshold value, the provided approximate curved lines are gathered as area information and processing is finished. Namely, when encoding the contour shapes of respective blocks, the approximate curved lines with a little generating information are used. |
