Computer aided tube and tip detection
A computer aided tube and tip detection method for a radiographic image. Radiographic image data is obtained, and a region of interest in the image is determined. The image is processed to provide edge enhancements forming an edge-enhanced image. Edge segments in the edge-enhanced image are detected. Connected lines from the edge segments are formed to obtain a set of connected lines. A tube structure is identified by pairing one or more pairs of connected lines that are separated by a width dimension in a predetermined range. A tip is detected for the tube structure according to the convergence or divergence of paired connected lines.
1. A method for processing a radiographic image, executed at least in part by a processor, comprising:
obtaining radiographic image data;
determining a region of interest in the image;
processing the image to provide edge enhancements and thereby form an edge-enhanced image;
detecting edge segments in the edge-enhanced image;
forming connected lines from the edge segments to form a set of connected lines;
identifying a tube structure by pairing one or more pairs of connected lines that are separated by a width dimension in a predetermined range; and
detecting a tip for the tube structure according to the convergence or divergence of paired connected lines.
2. The method of claim 1 wherein detecting edge segments comprises using Canny edge detection.
3. The method of claim 1 further comprising applying a Gaussian filter to the image data.
4. The method of claim 1 further comprising highlighting the tube structure and tip outline on the displayed radiographic image.
5. The method of claim 4 wherein highlighting the tube structure and tip outline comprises overlaying a color onto a displayed image.
6. The method of claim 1 wherein identifying the tube structure further comprises classifying the type of tube structure.
7. The method of claim 1 wherein detecting edge segments comprises applying a Hough transform.
8. The method of claim 1 further comprising dividing at least a portion of the image into a plurality of image tiles.
9. The method of claim 1 wherein detecting edge segments comprises computing a gradient.
10. The method of claim 1 wherein identifying a tube structure further comprises forming one or more additional connected pairs of lines to eliminate gaps in the tube structure.
11. The method of claim 1 further comprising performing histogram equalization to enhance grayscale contrast.
12. A method for processing a radiographic image, executed at least in part by a processor, comprising:
obtaining radiographic image data;
determining a region of interest in the image;
processing the image to provide edge enhancement, forming an edge-enhanced image thereby;
detecting edge segments in the edge-enhanced image;
pairing edge segments;
forming connected lines to join disconnected, paired edge segments; and
detecting a tip according to the convergence or divergence of paired connected lines,
wherein pairing edge segments further comprises:
obtaining gradient characteristics for the edge segments;
calculating the distance between paired edge segments; and
calculating the relative angles between edge segments.
13. The method of claim 12 further comprising identifying a type of tube according to the distance between paired segments.
14. A method for processing a radiographic image, executed at least in part by a processor, comprising:
obtaining radiographic image data;
determining a region of interest in the image;
processing the image to provide edge enhancement, forming an edge-enhanced image thereby;
detecting edge segments in the edge-enhanced image;
pairing edge segments;
forming connected lines to join disconnected, paired edge segments;
detecting a tip according to the convergence or divergence of paired connected lines; and
identifying a tube structure according to a width dimension that separates two paired, connected lines.
15. A method for processing a radiographic image, executed at least in part by a processor, comprising:
obtaining radiographic image data;
determining a region of interest in the image;
detecting edge segments in the region of interest;
forming connected lines from the edge segments to form a set of connected lines; and
identifying a tube structure by pairing one or more pairs of connected lines that are separated by a width dimension in a predetermined range.
16. The method of claim 15 further comprising detecting a tip for the tube structure according to the convergence or divergence of paired connected lines.