Image diagnosis supporting device

• US 20030179915A1
• Filed: 04/01/2003
• Published: 09/25/2003
• Est. Priority Date: 06/30/2000
• Status: Active Grant

• Current Assignee: FUJIFILM Healthcare Corp. (Fujifilm Holdings Corporation)
• Original Assignee: Hitachi Medical Corporation (Hitachi, Ltd.)
• Inventors: Goto, Yoshihiro
• Granted Patent: US 7,298,878 B2
• Time in Patent Office: Days
• US Class Current: 382/128
• CPC Class Codes: A61B 6/027 characterised by the use of a particular data acquisition trajectory, e.g. helical or spiral
• CPC Class Codes: A61B 6/463 characterised by displaying multiple images or images and diagnostic data on one display
• CPC Class Codes: A61B 6/468 allowing annotation or message recording
• CPC Class Codes: G06T 2200/24 involving graphical user interfaces [GUIs]
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/12 Edge-based segmentation

Abstract

An image diagnosis supporting device by which shadows of different sizes and shapes can be processed integrally with less computer calculation when focus candidates are automatically detected from a medical image. A medical image is converted to a multi-valued image by a digitizing means. The coordinates of the center of a shadow are determined according to the multi-valued image. With reference to the center coordinates, the medical image and the multi-valued image are subjected to various image-processings so as to detect portions considered as focus candidates. For example, a radius of predetermined length is rotated around a reference point

Claims

1. An image diagnosis supporting device characterized by:

digitizing means for applying predetermined image processing to a medical image and generating a multi-valued image comprising discrete multiple values;

extracting means for executing at least one decision making processing routine on the multi-valued image generated by the digitizing means and extracting from among shadows a focus candidate shadow, a shadow which is likely to indicate a diseased site; and

display means for displaying in the medical image the focus candidate shadow extracted by the extracting means so that it is easily identifiable.

2. An image diagnosis supporting device according to claim 1, characterized in that the extracting means extracts the focus candidate shadow from the multi-valued image generated by the digitizing means and the medical image.

3. An image diagnosis supporting device according to claim 1, further characterized in that image generating means for extracting from the medical image only pixels belonging to a pixel value range corresponding to the kind of shadow made by the target being searched and generating a medical image of the target, the digitizing means applying predetermined image processing to the target image and generating the multi-valued image, the extracting means extracting the focus candidate shadow from the multi-valued image and the decision target medical image on the basis of the multi-valued image generated by the digitizing means.

4. An image diagnosis supporting device according to claim 1, further characterized in that the extracting means adjusts a magnification ratio or a reduction ratio of the multi-valued image according to a size of a shadow which is a target, and extracts the focus candidate shadow from the multi-valued image that has been so adjusted.

5. An image diagnosis supporting device according to claim 1, characterized in that the extracting means selects a combination of at least one of the above decision making processing routines according to the slice thickness of the medical image, and extracts the focus candidate shadow from the multi-valued image through the selected combined processing routines.

6. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates on the shadow in the multi-valued image a radius of predetermined length about a reference point near the center or the weighted center of the shadow, samples pixels of the shadow in the multi-valued image which intersect the radius as it is rotated, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the pixel values.

7. An image diagnosis supporting device according to claim 6, characterized by sampling the values of pixels forming a spiral or forming concentric circles as the radius rotates, finding a representative value of each of loops formed by the rotation on the basis of the individual pixel values, comparing the representative value with a reference value stored in advance, and making a decision as to the nature of the shadow.

8. An image diagnosis supporting device according to claim 6, characterized by causing a radius of predetermined length to make rotations on each of the plurality of shadows in the multi-valued image about a reference point near a detection point of each of the shadows in the multi-valued image, sampling pixel values which intersect the radius, checking correlation of the pixel values of mutually adjacent loops formed by the rotations, and making a decision as to each of the shadows on the basis of the correlation.

9. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects the center or the weighted center of a shadow on the basis of the multi-valued image, rotates two straight lines which are at a predetermined angle η

to each other about a reference point near the center or the weighted center of a shadow in the multi-valued image, the medical image or the target medical image, samples pixel values of the shadow which intersect the two straight lines, searches for anisotropy of the shadow on the basis of the pixel values corresponding to the two straight lines, and makes a decision as to whether the shadow is a focus candidate shadow.

10. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, causes a radius of predetermined length to make rotations about a reference point near the center or the weighted center of a shadow in the multi-valued image, samples pixel values of the shadow where it intersects the radius, detects at least two radii at which the pixel values sharply change with the rotation and finds the angle of elevation, the angle between these radii, and compares the angle of elevation with a reference value stored in advance, and makes a decision as to whether the shadow is a focus candidate shadow.

11. An image diagnosis supporting device according to claim 10, characterized in that in the case where the decision making means compares the angle of elevation with the reference value stored in advance and determines that the shadow is a focus candidate shadow, the decision making means finds the length of contact between the shadow and a wall portion and determines on the basis of the length of contact whether the shadow is a focus candidate shadow or a shadow of an object accompanying a focus.

12. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a straight line of predetermined length about a reference point near the center or the weighted center of a shadow in the multi-valued image to find the maximum value or minimum value, or both, of a length of the portion of the straight line which intersects the shadow in the multi-valued image, samples first and second pixel values in the multi-valued image which are located a predetermined distance outward from the shadow along an extension of a straight line which passes through the reference point and is approximately perpendicular to a straight line with the above minimum value as well as third and fourth pixel values in the multi-valued image which are located a predetermined distance outward from the shadow along an extension of the straight line of the minimum value, fifth and sixth pixel values in the multi-valued image which are located a predetermined distance outward from the shadow along an extension of a straight line which passes through the reference point and is approximately perpendicular to the straight line with the above maximum value as well as seventh and eighth pixel values in the multi-valued image which are located a predetermined distance outward from the shadow along an extension of the straight line of the maximum value, or ninth to twelfth pixel values in the multi-valued image which are located a predetermined distance outward from the shadow along the extensions of the straight lines of the maximum value and the minimum value, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the first to fourth pixel values, the fifth to eighth pixel values or the ninth to twelfth pixel values.

13. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a radius of predetermined length about a reference point near the center or the weighted center of the shadow on the shadow in the multi-valued image to sample pixel values of the shadow which intersects the radius in the multi-valued image, generates a density waveform on the basis of the pixel values, finds the radii of at least two locations corresponding to angles at which peaks of the density waveform appear, finds the bisector of an angle made by the detected radii, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of the sum of the pixel values on the detected radii and the sum of the pixel values on the bisector.

14. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a radius of predetermined length about a reference point near the center or the weighted center of the shadow on the shadow in the multi-valued image to sample pixel values of the shadow which intersect the radius, generates a density waveform on the basis of the pixel values, finds radii lying at at least two locations at which peaks of the density waveform appear, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of the average value of the pixel values on the detected radii and the average value of the pixel values on radii other than the detected radii.

15. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a straight line of predetermined length about a reference point near the center or the weighted center of the shadow on the shadow to find the length of the portion of the straight line which intersects the shadow in the multi-valued image, performs Fourier expansion on a curve describing the length of the straight line portion with varying angle of rotation, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of a result of this Fourier expansion.

16. An image diagnosis supporting device according to claim 15, characterized by rotating the straight line of predetermined length on the shadow in the multi-valued image, plotting a curve where the vertical axis represents the length of the portion of the straight line which intersects the shadow and the horizontal axis represents the angle, detecting a positions of valleys of the curve, and performing interpolation between the values with a predetermined curve to draw the curve.

17. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a straight line of predetermined length about a reference point near the center or the weighted center of the shadow to find a minimum value of the length of the portion of the straight line which intersects the shadow, divides the area of the shadow by the second power of the minimum value, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the divided area.

18. An image diagnosis supporting device according to claim 1, characterized in that the extracting means detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a straight line of predetermined length about a reference point near the center or the weighted center of the shadow to find a minimum value of length of the portion of the straight line which intersects the shadow, finds a cutting length on the basis of the minimum value, and excludes an elongated shadow of the cutting length which touches the main shadow.

19. An image diagnosis supporting device according to claim 1, characterized in that the decision making means detects a center or a weighted center of a shadow on the basis of the multi-valued image, draws a closed curve corresponding to a shape of the shadow about a reference point near the center or the weighted center of the shadow, finds short curve lengths in a case where the closed curve is superposed on the shadow or short curve lengths in a case where the closed curve is not superposed on the shadow as well as the number of the short curve lengths, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of the relationship between the short curve lengths and the number of the short curve lengths.

20. An image diagnosis supporting device according to claim 1, characterized in that the decision making means samples pixel values in the medical image which intersect at least one straight line passing through the shadow and extending in a predetermined direction, find a positive or negative density gradient of each pixel on the straight line on the basis of the pixel values, defines the number of pixels in which the positive gradient continues as a positive run length and defines the number of pixels by which the negative gradient continues as a negative run length, finds positive and negative run lengths and the number of the positive run lengths and the number of positive run lengths, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of the relationship between the positive and negative run lengths and the number of the positive run lengths and the number of positive run lengths.

21. An image diagnosis supporting device according to claim 1, characterized in that the decision making means finds a variance or a standard deviation of pixel values of a shadow in the multi-valued image, and makes a decision as to whether the shadow is a focus candidate shadow on the basis of the variance or the standard deviation.

22. An image diagnosis supporting device according to claim 21, characterized by detecting a center or a weighted center of a shadow on the basis of the multi-valued image, rotating a straight line of predetermined length about a reference point near the center or the weighted center of the shadow to sample pixel values of the shadow which intersect the straight line in the multi-valued image, and finding the standard deviation of the pixel values which intersect the straight line at a predetermined angle, for making the decision as to the shadow.

23. An image diagnosis supporting device according to claim 21, characterized by dividing a shadow in the multi-valued image into a plurality of regions and finding a variance or a standard deviation of the pixel values in each of the regions, for making the decision as to the shadow.

24. An image diagnosis supporting device according to claim 21, characterized by finding an outside-shadow variance or an outside-shadow standard deviation of pixel values in a predetermined region outside a shadow in the multi-valued image, for making the decision as to the shadow.

25. An image diagnosis supporting device according to claim 21, characterized by detecting a center or a weighted center of a shadow in the multi-valued image and finding a variance or a standard deviation of distance from the center or the weighted center to an edge of the shadow along the entire periphery of the shadow, for making the decision as to the shadow.

26. An image diagnosis supporting device according to claim 21, characterized by finding a variance or a standard deviation of a distance from an edge to an edge of a shadow in the horizontal and vertical directions, for making the decision as to the shadow.

27. An image diagnosis supporting device according to claim 1, characterized in that the extracting means performs processing which detects a center or a weighted center of a shadow on the basis of the multi-valued image, rotates a straight line of predetermined length about a reference point near the center or the weighted center of a shadow on the multi-valued image, the medical image or the decision target medical image, finds a maximum value of length of the portion of the straight line which intersects the shadow, sets a strip-shaped extended line approximately parallel with a straight line of the maximum value, and adds a predetermined value to a pixel memory located on the strip-shaped extension, repeatedly executing this processing a number of times equal to the number of the shadows.

28. An image diagnosis supporting device according to claim 1, characterized in that the decision making means compares shadows existing in two multi-valued images adjacent to each other in a slice-thickness direction of the medical image, and makes a decision as to whether each of the shadows is a focus candidate shadow, on the basis of whether the shadows overlap each other in more than a predetermined proportion.

29. An image diagnosis supporting device according to claim 1, characterized in that the decision making means uses at least two sets from among a set of axial images, a set of sagittal images and a set of coronal images which are perpendicular to one another, and extracts a focus candidate shadow from each of the at least two sets and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the position of the focus candidate shadow extracted in each of the images.

30. An image diagnosis supporting device according to claim 29, characterized by, in the case where the shadow is a focus candidate shadow, storing in a memory the coordinate position of the focus candidate shadow and information on the focus candidate shadow.

31. An image diagnosis supporting device according to claim 1, characterized in that the decision making means finds the area of the shadow region and also finds an area of a concave region formed in an edge portion of the shadow region, finds a ratio of the area of the shadow region to the area of the concave region, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the found ratio.

32. An image diagnosis supporting device according to claim 1, characterized in that the decision making means finds an area of a circle or a polygon inscribed in an edge of the shadow region and the area of separated regions into which the shadow is divided by the circle or the polygon, and finds the ratio of the area of the circle or the polygon to the area of the separated regions of the shadow, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the found ratio.

33. An image diagnosis supporting device according to claim 1, characterized in that the decision making means finds an area of a circle or a polygon circumscribed to the outer edge of the shadow region, finds a ratio of the area of the circle to the area of the shadow, and makes a decision as to whether the shadow is a focus candidate shadow, on the basis of the found ratio.

34. An image diagnosis supporting device according to claim 1, characterized in that the display means displays a shadow determined as the focus candidate shadow through the decision making processing routines in the medical image or in an area other than the medical image each time one processing of the decision making processing routines is completed.

35. An image diagnosis supporting device according to claim 34, characterized in that a first display area for displaying a focus candidate shadow or image supplementary information, a second display area for displaying an undetected image in which a focus candidate is not detected or image supplementary information, and a third display area for displaying an image impossible to identify and image supplementary information are provided on a picture of the display means.

36. An image diagnosis supporting device according to claim 1, characterized in that the display means displays a magnified image of a vicinity of the focus candidate shadow in the medical image or in an area other than the medical image.

37. An image diagnosis supporting device according to claim 1, characterized in that the display means displays the medical image by controlling the order of display thereof according to the position of the focus candidate shadow in the medical image.

38. An image diagnosis supporting device according to claim 1, characterized in that the display means displays the medical images having the focus candidate shadow in an order controlled by a pointing device, according to the movement the pointing device in the medical image with the focus candidate shadow.

39. An image diagnosis supporting device according to claim 1, characterized in that the display means displays a marker line which encloses the extracted focus candidate shadow.

40. An image diagnosis supporting device according to claim 39, characterized in that the extracting means calculates the probability that a shadow is a focus, and the display means changes displays the marker line in the display on the basis of the focus certainty.

41. An image diagnosis supporting device according to claim 39, characterized in that in the case where markers overlap to enclose a plurality of focus candidate shadows, the display means displays the markers with an overlapping portion thereof erased.

42. An image diagnosis supporting device according to claim 39, characterized in that the display means performs contrast emphasis processing or gamma emphasis processing on the area enclosed with the marker and clearly displays the focus candidate shadow.

43. An image diagnosis supporting device according to claim 39, characterized in that in the case where a hidden mode in which display of the marker is disabled is selected, the display means stops displaying the marker and displays, on a picture, information indicating that the hidden mode is presently active, and automatically displays the marker when a predetermined time elapses after the hidden mode has been started.

44. An image diagnosis supporting device according to claim 1, characterized in that the extracting means finds length of contact between the extracted focus candidate shadow and a wall portion and determines on the basis of the length of contact whether the shadow is a shadow of an object accompanying a focus, the display means displaying the shadow of an object accompanying a focus enclosed with a marker.

45. An image diagnosis supporting device according to claim 1, characterized in that the display means displays focus candidate shadows respectively extracted from medical images photographed at mutually different points of time, with the respective focus candidate shadows enclosed with markers which makes it possible to discriminate between the points of time of photography.

46. An image diagnosis supporting device according to claim 1, characterized in that the display means displays a marker having an elliptical shape whose long-axis direction coincides with a long-axis direction of the extracted focus candidate shadow, with the focus candidate shadow enclosed with the marker.

47. An image diagnosis supporting device according to claim 46, characterized in that in the case where markers overlap one another to enclose a plurality of focus candidate shadows, respectively, the display means displays the markers with the overlapping portion thereof erased.

48. An image diagnosis supporting device according to claim 46, characterized in that the display means performs contrast emphasis processing or gamma emphasis processing on an area enclosed with the markers and clearly displays the focus candidate shadows.

49. An image diagnosis supporting device according to claim 46, characterized in that in the case where a hidden mode in which display of the markers is disabled is selected, the display means stops displaying the markers and displays information indicating that the hidden mode is presently active, and automatically displays the markers when a predetermined time elapses after the hidden mode has been started.

50. An image diagnosis supporting device according to claim 1, characterized in that the display means displays a medical image in which the focus candidate shadow exists and a medical image in which the focus candidate shadow does not exist, by moving image display with different display times allocated for the respective medical images.

51. An image diagnosis supporting device according to claim 1, characterized in that the display means does not display a medical image which has not yet undergone shadow detection by a doctor so that it can be identified.