Image processing device, computer readable storage medium storing image processing program, and image processing method

1. An image processing device comprising:

a feature value calculating unit that calculates a color feature value of each pixel or each segment inside at least one of a series of images captured by an image pickup device while moving inside a gastrointestinal tract;

a color feature value estimating unit that, based on data distribution of the color feature value of each pixel or each segment, estimates color feature value data representing mucous membrane and color feature value data representing gastrointestinal contents;

a discrimination boundary setting unit that, based on the estimated color feature value data, sets a color feature value discrimination boundary between mucous membrane and gastrointestinal contents;

a mucosal area identifying unit that identifies a mucosal area in at least one of the series of intra-gastrointestinal-tract images, using the set color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents; and

an image selecting unit that selects a plurality of images from the series of images captured by the image pickup device while moving inside the gastrointestinal tract, wherein

the feature value calculating unit calculates the color feature value of each pixel or each segment inside the selected plurality of images.

2. The image processing device according to claim 1 , wherein the color feature value estimating unit includes a clustering unit for dividing the color feature value data into clusters, estimates that a cluster having maximum count of color feature value data is the color feature value data representing mucous membrane, and estimates that a cluster that has a relative position thereof with the cluster having maximum count of color feature value data within a predetermined range in a color feature space is the color feature value data representing gastrointestinal contents.

3. The image processing device according to claim 1 , wherein the discrimination boundary setting unit includes a bimodal determining unit for determining bimodal of a histogram in a color feature space of the color feature value data estimated by the color feature value estimating unit, performs setting processing according to a determination result of the bimodal determining unit, and sets the color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents.

4. The image processing device according to claim 3 , wherein, when the bimodal determining unit determines that the histogram has a high bimodal, then the discrimination boundary setting unit performs, as the setting processing, setting of a position of a valley in the histogram as the color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents.

5. The image processing device according to claim 3 , wherein, when the bimodal determining unit determines that the histogram has a low bimodal, then the discrimination boundary setting unit performs, as the setting processing, setting of the color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents based on a representative value of the histogram and information on variation in color feature value data corresponding to the representative value.

6. The image processing device according to claim 1 , wherein the color feature value estimating unit estimates that color feature value data representing intestinal mucosa is the color feature value data representing mucous membrane.

7. The image processing device according to claim 1 , wherein the image selecting unit selects, from the series of images, a plurality of images from among images captured after an average transit time taken by the image pickup device to pass through stomach.

8. The image processing device according to claim 1 , wherein the image selecting unit selects, from the series of images, a plurality of images from among images captured before an average transit time taken by the image pickup device to pass through small intestine.

9. The image processing device according to claim 1 , wherein the color feature value is hue.

10. The image processing device according to claim 1 , wherein the feature value calculating unit includes a segment partitioning unit for partitioning each image of the series of images captured inside the gastrointestinal tract into segments and calculates a color feature value of each segment partitioned by the segment partitioning unit.

11. A non-transitory computer readable storage medium having stored therein an image processing program, the image processing program comprising instructions for causing a computer to execute:

calculating a color feature value of each pixel or each segment inside at least one of a series of images captured by an image pickup device while moving inside a gastrointestinal tract;

estimating, based on data distribution of the color feature value of each pixel or each segment, color feature value data representing mucous membrane and color feature value data representing gastrointestinal contents;

setting, based on the estimated color feature value data, a color feature value discrimination boundary between mucous membrane and gastrointestinal contents; and

identifying a mucosal area in at least one of the series of intra-gastrointestinal-tract images, using the set color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents; and

selecting a plurality of images from the series of images captured by the image pickup device while moving inside the gastrointestinal tract, wherein

the calculating a color feature value comprises calculating the color feature value of each pixel or each segment inside the selected plurality of images.

12. An image processing method comprising:

calculating, using an image processing device, a color feature value of each pixel or each segment inside at least one of a series of images captured by an image pickup device while moving inside a gastrointestinal tract;

estimating, based on data distribution of the color feature value of each pixel or each segment, color feature value data representing mucous membrane and color feature value data representing gastrointestinal contents;

setting, based on estimated color feature value data, a color feature value discrimination boundary between mucous membrane and gastrointestinal contents; and

identifying a mucosal area in at least one of the series of intra-gastrointestinal-tract images, using the set color feature value discrimination boundary between the mucous membrane and the gastrointestinal contents; and

selecting a plurality of images from the series of images captured by the image pickup device while moving inside the gastrointestinal tract, wherein

the calculating a color feature value comprises calculating the color feature value of each pixel or each segment inside the selected plurality of images.