Image processing apparatus, image processing method, and computer-readable recording medium

1. An image processing apparatus for processing an image obtained by capturing inside of a lumen of a living body, the image processing apparatus comprising:

a processor comprising hardware, wherein the processor is configured to:

perform a contour edge region extracting process comprising extracting a contour edge region of an examination target from the image;

perform an examination region setting process comprising setting an examination region in the image so as not to include the contour edge region; and

perform an abnormal structure identifying process comprising identifying whether a microstructure of a surface of the examination target is abnormal, based on texture information of the examination region,

wherein the examination region setting process comprises:

performing a candidate region setting process comprising setting examination candidate regions in the image; and

performing a region determining process comprising determining the examination region from the examination candidate regions, based on information on the contour edge region, and

wherein the region determining process comprises:

performing a contour edge region eliminating process comprising eliminating the contour edge region present in the examination candidate regions; and

determining, as the examination region, an examination candidate region that remains after the contour edge region is eliminated by the contour edge region eliminating process.

2. The image processing apparatus according to claim 1 ,

wherein the candidate region setting process comprises:

performing a representative-imaging-distance computing process comprising computing a representative imaging distance to the examination target at a position where each of the examination candidate regions is to be set, and

wherein the candidate region setting process comprises setting the examination candidate regions each having a size according to the representative imaging distance.

3. The image processing apparatus according to claim 1 ,

wherein the candidate region setting process comprises:

performing a representative-imaging-distance-gradient computing process comprising computing a representative imaging distance gradient of the examination target at a position where each of the examination candidate regions is to be set, and

wherein the candidate region setting process comprises setting the examination candidate regions each having a size according to the representative imaging distance gradient.

4. The image processing apparatus according to claim 1 ,

wherein the examination region setting process comprises:

performing a distance transform image computing process comprising computing a distance transform image where a distance between each pixel in the image and the contour edge region or an edge of the image is transformed into a pixel value, and

wherein the examination region setting process comprises setting the examination region based on the distance transform image.

5. The image processing apparatus according to claim 4 ,

wherein the examination region setting process comprises setting the examination region such that a pixel whose pixel value is a maximum value in the distance transform image is at a center of the examination region and a distance from the center to a boundary of the examination region is smaller than a distance corresponding to the maximum value.

6. The image processing apparatus according to claim 1 ,

wherein the processor is further configured to:

perform a repetition process comprising repeatedly performing, on regions where the examination region has not been set, the examination region setting process.

7. The image processing apparatus according to claim 6 ,

wherein the repetition process comprises changing contents of the examination region setting process, according to repetition of the examination region setting process.

8. The image processing apparatus according to claim 1 ,

wherein the contour edge region extracting process comprises:

performing an edge strength computing process comprising computing an edge strength of each pixel in the image;

performing a contour edge candidate region extracting process comprising extracting, as a contour edge candidate region, a pixel with the edge strength greater than or equal to a predetermined threshold value; and

performing a contour edge region specifying process comprising specifying the contour edge region, based on feature data of the contour edge candidate region.

9. The image processing apparatus according to claim 1 ,

wherein the image includes a plurality of wavelength components corresponding to a plurality of colors,

wherein the contour edge region extracting process comprises:

performing a low-absorption wavelength selecting process comprising selecting a low-absorption wavelength component among the plurality of wavelength components, the low-absorption wavelength component being a wavelength component with a lowest degree of absorption or scattering in the living body, and

wherein the contour edge region extracting process comprises extracting the contour edge region based on the low-absorption wavelength component.

10. The image processing apparatus according to claim 1 ,

wherein the abnormal structure identifying process comprises:

performing a specific frequency component computing process comprising computing a specific spatial frequency component in the examination region; and

performing a statistical classification process comprising performing statistical classification based on the specific spatial frequency component.

11. The image processing apparatus according to claim 10 ,

wherein the image includes a plurality of wavelength components corresponding to a plurality of colors, and

wherein the abnormal structure identifying process comprises:

performing a specific wavelength component selecting process comprising selecting a specific wavelength component from the plurality of wavelength components, the specific wavelength component being specified according to a degree of absorption or scattering in the living body.

12. The image processing apparatus according to claim 10 ,

wherein the image includes a plurality of wavelength components corresponding to a plurality of colors, and

wherein the abnormal structure identifying process comprises:

performing a ratio-between-specific-wavelengths computing process comprising computing a ratio between wavelength components having different degrees of absorption or scattering in the living body.

13. The image processing apparatus according to claim 10 ,

wherein the abnormal structure identifying process comprises:

performing a representative-imaging-distance acquiring process comprising acquiring a representative imaging distance to the examination target in the examination region, and

wherein the specific frequency component computing process comprises computing the specific spatial frequency component specified according to the representative imaging distance.

14. The image processing apparatus according to claim 1 ,

wherein the processor is further configured to perform:

a non-examination target region eliminating process comprising eliminating a non-examination target region included in the image from setting of the examination region.

15. The image processing apparatus according to claim 14 ,

wherein the non-examination target region is a region where one of a dark part, a bright part, a residue, and a bubble is present.

16. An image processing method executed by a processor comprising hardware for processing an image obtained by capturing inside of a lumen of a living body, the image processing method comprising:

a contour edge region extracting step of extracting a contour edge region of an examination target from the image;

an examination region setting step of setting an examination region in the image so as not to include the contour edge region; and

an abnormal structure identifying step of identifying whether a microstructure of a surface of the examination target is abnormal, based on texture information of the examination region,

wherein the examination region setting step comprises:

a candidate region setting step of setting examination candidate regions in the image; and

a region determining step of determining the examination region from the examination candidate regions, based on information on the contour edge region, and

wherein the region determining step comprises:

a contour edge region eliminating step comprising eliminating the contour edge region present in the examination candidate regions; and

determining, as the examination region, an examination candidate region that remains after the contour edge region is eliminated by the contour edge region eliminating step.

17. A non-transitory computer-readable recording medium with an executable program stored thereon, the program instructing a processor comprising hardware for processing an image obtained by capturing inside of a lumen of a living body, to execute:

a contour edge region extracting step of extracting a contour edge region of an examination target from the image;

an examination region setting step of setting an examination region in the image so as not to include the contour edge region; and

an abnormal structure identifying step of identifying whether a microstructure of a surface of the examination target is abnormal, based on texture information of the examination region,

wherein the examination region setting step comprises:

a candidate region setting step of setting examination candidate regions in the image; and

a region determining step of determining the examination region from the examination candidate regions, based on information on the contour edge region, and

wherein the region determining step comprises:

a contour edge region eliminating step comprising eliminating the contour edge region present in the examination candidate regions; and

determining, as the examination region, an examination candidate region that remains after the contour edge region is eliminated by the contour edge region eliminating step.