| 摘要 |
Based on a long diameter of a set specific region, a first region estimated as the specific region and a second region estimated as a background region are set within an input image. Based on a density histogram of each pixel in the first region and a density histogram of each pixel in the second region, a first evaluation value which indicates a likelihood that a density value represents the specific region is calculated for each density value. For each pair of two adjacent pixels within the input image, based on the first evaluation values, a second evaluation value which indicates a likelihood that the two adjacent pixels represent a contour is calculated. Based on the second evaluation value, connection costs between a plurality of nodes respectively representing each pair of the two adjacent pixels are set. The plurality of nodes are connected based on the connection costs. |
| 主权项 |
1. An image processing device, comprising:
a memory configured to store processing instructions; and a processor configured to execute the stored processing instructions, which comprise: setting a long diameter of a specific region within an input image containing the specific region; setting, based on the set long diameter, a region of a predetermined shape including the set long diameter as a first region estimated as highly probable to be the specific region within the input image and a region of a predetermined shape outside the first region as a second region estimated as highly probable to be a background region, which is a region other than the specific region within the input image; calculating a first histogram, which is a density histogram of a density value of each pixel in the first region, and a second histogram, which is a density histogram of a density value of each pixel in the second region, calculating, for each density value, a first evaluation value which indicates a likelihood that the each density value represents the specific region based on the first histogram and the second histogram, and calculating, for each pair of two adjacent pixels within the input image, a second evaluation value which indicates a likelihood of being a contour, which is a likelihood that the two adjacent pixels represent the contour, based on the first evaluation values respectively corresponding to the density values of the two adjacent pixels; and extracting the specific region by defining a first reference node belonging to the specific region, a second reference node belonging to the background region, and a plurality of nodes representing respective pixels within the input image, setting a connection cost indicating how likely the respective nodes are to be connected to each other based on the second evaluation value, and connecting the respective nodes to each other based on the set connection cost. |
