Method and apparatus for classifying detection inputs in medical images

• US 20070177781A1
• Filed: 01/31/2006
• Published: 08/02/2007
• Est. Priority Date: 01/31/2006
• Status: Active Grant

• Current Assignee: MeVis Medical Solutions AG (Varex Imaging Corporation)
• Original Assignee: MeVis Medical Solutions, Inc. (Varex Imaging Corporation)
• Inventors: Raffy, Philippe
• Granted Patent: US 7,623,694 B2
• Time in Patent Office: Days
• US Class Current: 382/128
• CPC Class Codes: G06F 18/2415 based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images
• CPC Class Codes: G06V 30/19173 Classification techniques
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches

Abstract

Detection inputs are classified using two thresholds. In a preferred embodiment, a multiplicity of inputs are scored and the scored inputs are searched to locate an input that has a score greater than a first threshold. If such input is found, every input having a score in excess of a second threshold lower than the first threshold is identified as belonging to a first class of interest. If no input is found having a score in excess of the first threshold, no inputs are identified as belonging to the first class.

Claims

1. A method of classifying detection inputs as belonging to a first class or a second class depending on a numerical score associated with each input comprising the steps of:

scoring a multiplicity of detection inputs to produce a set of scores;

comparing each of the scores with a first threshold,

if at least one score exceeds the first threshold, classifying in the first class all inputs having a score that exceeds a second threshold lower than the first threshold, and

if no score exceeds the first threshold, classifying none of the inputs in the first class.

2. The method of claim 1 wherein the detection inputs are for pulmonary embolisms.

3. The method of claim 1 further comprising the step of setting a first threshold that is the score of a known true positive input in a training set used to train an algorithm for scoring the detection inputs.

4. The method of claim 1 wherein the first threshold is set at the lowest score of a group of highest scores for known true positive inputs in a training set used to train an algorithm for scoring the detection inputs.

5. The method of claim 1 further comprising the step of training an algorithm for scoring the detection inputs.

6. A method of classifying detection inputs as belonging to a first class or a second class depending on a numerical score associated with each input comprising the steps of:

scoring a multiplicity of detection inputs to produce a set of scores;

comparing each of the scores with a first threshold,

if at least one score is less than the first threshold, classifying in the first class all inputs having a score that is less than a second threshold greater than the first threshold, and

if no score is less than the first threshold, classifying none of the inputs in the first class.

7. The method of claim 6 wherein the detection inputs are pulmonary embolisms.

8. The method of claim 6 further comprising the step of setting a first threshold that is the score of a known true positive input in a training set used to train an algorithm for scoring the detection inputs.

9. The method of claim 6 wherein the first threshold is set at the highest score of a group of lowest scores for known true positive inputs in a training set used to train an algorithm for scoring the detection inputs.

10. The method of claim 6 further comprising the step of training an algorithm for scoring the detection inputs.

11. A method of classifying detection inputs as belonging to a first class or a second class depending on a numerical score associated with each input comprising the steps of:

scoring a multiplicity of detection inputs to produce a set of scores;

comparing each of the scores with a first threshold,

if at least one score exceeds the first threshold, lowering a second threshold that is lower than the first threshold,

if no score exceeds the first threshold, raising the second threshold, and

classifying in the first class all inputs having a score that exceeds the second threshold.

12. The method of claim 11 wherein the detection events are for pulmonary embolisms.

13. The method of claim 11 further comprising the step of setting a first threshold that is the score of a known true positive input in a training set used to train an algorithm for scoring the detection inputs.

14. The method of claim 11 wherein the first threshold is set at the lowest score of a group of highest scores for known true positive inputs in a training set used to train an algorithm for scoring the detection inputs.

15. The method of claim 11 further comprising the step of training an algorithm for scoring the detection inputs.

16. A method of classifying detection inputs as belonging to a first class or a second class depending on a numerical score associated with each input comprising the steps of:

scoring a multiplicity of detection inputs to produce a set of scores;

comparing each of the scores with a first threshold,

if at least one score is less than the first threshold, raising a second threshold that is greater than the first threshold,

if no score is less than the first threshold, lowering the second threshold, and

classifying in the first class all inputs having a score that are less than a second threshold.

17. The method of claim 16 wherein the detection events are for pulmonary embolisms.

18. The method of claim 16 further comprising the step of setting a first threshold that is the score of a known true positive input in a training set used to train an algorithm for scoring the detection inputs.

19. The method of claim 16 wherein the first threshold is set at the highest score of a group of lowest scores for known true positive inputs in a training set used to train an algorithm for scoring the detection inputs.

20. The method of claim 16 further comprising the step of training an algorithm for scoring the detection inputs.

21. A computer program stored in a computer-readable medium for detecting abnormalities in medical images said program comprising instructions for classifying inputs as belonging to a first class or a second class depending on a numerical score associated with each input, said instructions comprising:

scoring a multiplicity of detection inputs to produce a set of scores;

comparing each of the scores with a first threshold,

if at least one score is greater than the first threshold, classifying in the first class all inputs having a score that is greater a second threshold less stringent than the first threshold, and

if no score is greater than the first threshold, classifying none of the inputs in the first class.

22. The computer program of claim 21 wherein the detection inputs are for pulmonary embolisms.