Film to DICOM conversion

• US 9,965,588 B2
• Filed: 03/06/2014
• Issued: 05/08/2018
• Est. Priority Date: 03/06/2014
• Status: Active Grant

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Sankhe, Pritesh, Gupta, Mithun Das
• Primary Examiner(s): Liew, Alex
• Application Number:
  US14/199,086
• Publication Number:
  US 20150254401A1
• Time in Patent Office: 1,524 Days
• Field of Search: 382128-134, 382181-231
• CPC Class Codes: G16H 30/20 for handling medical images, e.g. DICOM, HL7 or PACS
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: H04N 1/00127 Connection or combination of a still picture apparatus with another apparatus, e.g. for storage, processing or transmission of still picture signals or of information associated with a still picture
• CPC Class Codes: H04N 2201/3247 Data linking a set of images to one another, e.g. sequence, burst or continuous capture mode

Abstract

Embodiments digitize radiology films into DICOM format. Radiology films typically include an array of captured images laid out in a grid pattern. To comply with DICOM format, the scanned image of a radiology film is segmented into sub-images and text is extracted from the sub-images to generate DICOM metadata. The sub-images and extracted text metadata are then combined to generate a DICOM-compliant multi-image file.

Claims

1. A computer-implemented method comprising:

performing a Hough Transform to detect straight lines within an image of a scanned radiology film, the image including a plurality of sub-images in a grid pattern;

discarding false positives from the detected straight lines based on a statistical analysis of lengths of the detected straight lines;

identifying gridlines between sub-images based on a clustering of remaining detected straight lines;

separating sub-images based on the gridlines; and

storing the separated sub-images.

2. The method of claim 1, wherein discarding the false positives from the detected straight lines comprises:

finding a Simple Moving Average of the lengths of the detected straight lines sorted by distance from an origin;

determining a mean and standard deviation of a distribution of the Simple Moving Average; and

discarding a line as a false positive responsive to the Simple Moving Average corresponding to the line being less than a cutoff number of standard deviations from the mean.

3. The method of claim 1, further comprising detecting text regions within the sub-images by:

detecting horizontal and vertical edges of the text regions by applying Sobel operators along an X direction and along a Y direction;

connecting the detected horizontal and vertical edges into clusters by performing morphological dilation along the X direction and along the Y direction; and

performing a binary AND operation to identify text regions that have horizontal and vertical edges within a same region of a respective sub-image.

4. The method of claim 1, further comprising:

detecting text regions within the sub-images;

applying Minimum Cross Entropy thresholding to detected text regions; and

performing hole filling on the detected text regions.

5. The method of claim 1, further comprising:

detecting text regions within the sub-images;

performing optical character recognition to recognize text from a respective text region of a respective sub-image; and

storing the text associated with the respective sub-image.

6. The method of claim 5, further comprising:

accessing the separated sub-images and text associated with the respect sub-images;

formatting a DICOM-compliant multi-image file based on the sub-images and associated text; and

outputting the DICOM-compliant file.

17. The method of claim 1, wherein identifying the gridlines between sub-images based on the clustering of the remaining detected straight lines comprises:

clustering remaining detected straight lines within a short distance of each other; and

identifying a mean position of each cluster as a position of a gridline.

18. The method of claim 1, wherein identifying gridlines between sub-images based on the clustering of the remaining detected straight lines comprises:

determining a common repeating distance between gridlines identified by the clustering; and

identifying missing gridlines by extrapolating according to the common repeating distance from the gridlines identified by the clustering.

7. A non-transitory computer-readable storage medium storing executable computer program instructions, the instructions executable to perform steps comprising:

performing a Hough Transform to detect straight lines within an image of a scanned radiology film, the image including a plurality of sub-images in a grid pattern;

discarding false positives from the detected straight lines based on a statistical analysis of lengths of the detected straight lines;

identifying gridlines between sub-images based on a clustering of remaining detected straight lines;

separating sub-images based on the gridlines; and

storing the separated sub-images.

8. The medium of claim 7, wherein discarding the false positives from the detected straight lines comprises:

finding a Simple Moving Average of the lengths of the detected straight lines sorted by distance from an origin;

determining a mean and standard deviation of a distribution of the Simple Moving Average; and

discarding a line as a false positive responsive to the Simple Moving Average corresponding to the line being less than a cutoff number of standard deviations from the mean.

9. The medium of claim 7, further comprising detecting text regions within the sub-images by:

detecting horizontal and vertical edges of the text regions by applying Sobel operators along an X direction and along a Y direction;

connecting the detected horizontal and vertical edges into clusters by performing morphological dilation along the X direction and along the Y direction; and

performing a binary AND operation to identify text regions that have horizontal and vertical edges within a same region of a respective sub-image.

10. The medium of claim 7, the steps further comprising:

detecting text regions within the sub-images;

applying Minimum Cross Entropy thresholding to detected text regions; and

performing hole filling on the detected text regions.

11. The medium of claim 7, the steps further comprising:

detecting text regions within the sub-images;

performing optical character recognition to recognize text from a respective text region of a respective sub-image; and

storing the text associated with the respective sub-image.

12. The medium of claim 11, the steps further comprising:

accessing the separated sub-images and text associated with the respect sub-images;

formatting a DICOM-compliant multi-image file based on the sub-images and associated text; and

outputting the DICOM-compliant file.

19. The medium of claim 7, wherein identifying the gridlines between sub-images based on the clustering of the remaining detected straight lines comprises:

clustering remaining detected straight lines within a short distance of each other; and

identifying a mean position of each cluster as a position of a gridline.

20. The medium of claim 7, wherein identifying gridlines between sub-images based on the clustering of the remaining detected straight lines comprises:

determining a common repeating distance between gridlines identified by the clustering; and

identifying missing gridlines by extrapolating according to the common repeating distance from the gridlines identified by the clustering.

13. A system comprising:

a processor configured to execute modules; and

a memory storing the modules, the modules comprising;

an image segmentation module comprising;

a straight-line detector configured to

perform a Hough Transform to detect straight lines within an image of a scanned radiology film, the image including a plurality of sub-images in a grid pattern,

discard false positives from the detected straight lines based on a statistical analysis of lengths of the detected straight lines, and

identify gridlines between sub-images based on a clustering of remaining detected straight lines;

a sub-image separator configured to separate sub-images based on the gridlines; and

a sub-image storage configured to store the separated sub-images.

14. The system of claim 13, wherein the modules further comprise a metadata extraction module comprising a text region detection module configured to detect text regions within the sub-images.

15. The system of claim 14, wherein the metadata extraction module further comprises:

a text region processing module; and

a text recognition module configured to recognize text from text regions of respective sub-images.

16. The system of claim 15, wherein the modules further comprise a DICOM file generation module configured to format a DICOM-compliant multi-image file based on the sub-images and the recognized text.