Method for the automatic recognition of an object in an image

• US 7,860,281 B2
• Filed: 08/31/2006
• Issued: 12/28/2010
• Est. Priority Date: 09/01/2005
• Status: Active Grant

• Current Assignee: Siemens Healthcare GMBH (Siemens AG)
• Original Assignee: Siemens AG
• Inventors: Pfister, Marcus, Sandkamp, Bernhard
• Primary Examiner(s): Johns; Andrew W
• Assistant Examiner(s): ANSARI, TAHMINA N
• Application Number:
  US11/515,285
• Publication Number:
  US 20070047769A1
• Time in Patent Office: 1,580 Days
• Field of Search: 382/100, 382/128, 382/130, 382/131, 382/132, 382/133, 382/134, 382/155, 382/156, 382/157, 382/158, 382/159, 382/160, 382/162, 382/165, 382/173, 382/181, 382/190, 382/224, 382/227, 600/300, 128/920, 128/922, 128/923, 128/924, 128/925, 706/10, 706/11, 706/12, 706/13, 706/14, 706/15, 706/16, 706/17, 706/18, 706/19, 706/20, 706/21, 706/25, 700/48, 700/49
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/10121 Fluoroscopy
• CPC Class Codes: G06T 2207/20132 Image cropping
• CPC Class Codes: G06T 2207/30021 Catheter; Guide wire
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/74 involving reference images or patches

Abstract

To enable a rapid and safe automatic recognition of the position of the catheter, particularly for an electrophysiological examination with an automatic control system of an electrode catheter, a two-stage image evaluation of an x-ray fluoroscopy image is provided. In the first analysis step, frequency polygons are herewith developed as potential locations of the catheter with the aid of structures contained in the image. In the second analysis step, only the previously identified potential catheter positions are investigated with the aid of a template method and the actual catheter position is identified.

Claims

1. A method for automatically recognizing an object in an image obtained by a medical imaging method, comprising:

determining with a processor a plurality of potential object locations located over an entire area of the image, wherein the determining of the plurality of potential object locations is based on a first analysis method for image evaluation and for recognition of respective structures and patterns at the plurality of potential object locations; and

investigating with the processor the plurality of potential object locations for identifying an actual object location, wherein the investigating of the plurality of potential object locations is based on a second analysis method for image evaluation and for recognition of structures and patterns at selected areas of the image, wherein the selected areas of the image being evaluated by the second analysis method correspond to the plurality of potential object locations determined by the first analysis method,

wherein the first analysis method has an image evaluation accuracy corresponding to a degree of probability that structures and patterns recognized by the first analysis method at the plurality of potential object locations actually correspond to the object;

determining with the first analysis method a plurality of frequency polygons indicative of the plurality of potential object locations;

comparing each of the plurality of frequency polygons relative to a predetermined frequency polygon assigned to the object;

selecting frequency polygons which have a resemblance to the predetermined object frequency polygon as likely potential object locations;

wherein the second analysis method has an image evaluation accuracy corresponding to a degree of probability that structures and patterns recognized by the second analysis method at the plurality of potential object locations actually correspond to the object, wherein the second analysis method involves a template or map of the object;

laying the template or map of the object over the potential object locations in the image;

comparing the template or map of the object relative to the frequency polygons corresponding the potential object location;

based on the results of the comparing of the template or the map relative to the frequency polygons, selecting one location likely to correspond to the actual location of the object;

wherein the degree of probability of the second analysis method is higher relative to the degree of probability of the first analysis method, and

wherein the image evaluation accuracy of the first analysis method is adjustable by a user to adjust a number of the plurality of frequency polygons being determined by the first analysis method.

2. The method as claimed in claim 1, wherein the actual object location is determined continuously in parallel to a generation of the image.

3. The method as claimed in claim 2, wherein the actual object location is used for an automatic navigation of the object.

4. The method as claimed in claim 1, wherein the object is a medical instrument.

5. The method as claimed in claim 4, wherein the medical instrument is a catheter.

6. A computer for automatically recognizing an object in an image obtained by a medical imaging method, comprising:

a monitor that displays the image; and

a central processing unit that;

determines a plurality of potential object locations located over an entire area of the image, wherein the determining of the plurality of potential object locations is based on a first analysis method for image evaluation and for recognition of respective structures and patterns at the plurality of potential object locations;

investigates the plurality of potential object locations for identifying an actual object location, wherein the investigating of the plurality of potential object locations is based on a second analysis method for image evaluation and for recognition of structures and patterns at selected areas of the image, wherein the selected areas of the image being evaluated by the second analysis method correspond to the plurality of potential object locations determined by the first analysis method,

wherein the first analysis method has an image evaluation accuracy corresponding to a degree of probability that structures and patterns recognized by the first analysis method at the plurality of potential object locations actually correspond to the object;

determines with the first analysis method a plurality of frequency polygons;

compares each of the plurality of frequency polygons relative to a predetermined frequency polygon assigned to the object;

selects frequency polygons which have a resemblance to the predetermined object frequency polygon as potential object locations;

wherein the second analysis method has an image evaluation accuracy corresponding to a degree of probability that structures and patterns recognized by the second analysis method at the plurality of potential object locations actually correspond to the object, wherein the second analysis method involves a template or map of the object;

lays the template or map of the object over the potential object locations in the image;

compares the template or map of the object relative to the frequency polygons corresponding the potential object locations;

based on the results of the comparison of the template or map relative to the frequency polygons, selects one location likely to correspond to the actual location of the object;

wherein the degree of probability of the second analysis method is higher relative to the degree of probability of the first analysis method, and

wherein the image evaluation accuracy of the first analysis method is adjustable by a user to adjust a number of the plurality of frequency polygons being determined by the first analysis method.

7. The computer as claimed in claim 6, wherein the central processing unit comprises a software which executes the determining and investigating steps.