Automatic summarization of medical imaging studies

1. A method, in a data processing system comprising a processor and a memory, the memory comprising instructions that are executed by the processor to configure the processor to implement a medical imaging study summary engine, the method comprising:

receiving, by the medical imaging study summary engine, a medical imaging study data structure having data representing a plurality of medical images of a patient;

generating, by the medical imaging study summary engine, a temporal trajectory data structure of at least a subset of the medical images in the plurality of medical images, wherein the temporal trajectory data structure specifies topological changes in temporally subsequent medical images in the plurality of medical images;

selecting, by the medical imaging study summary engine, medical image data corresponding to selected medical images from the medical imaging study data structure based on the temporal trajectory data structure; and

outputting, by the medical imaging study summary engine, the selected medical image data via a medical imaging study user interface at least by generating a storyboard representation of the medical imaging study based on the selected medical images.

2. The method of claim 1 , wherein generating the temporal trajectory data structure comprises, for each medical image in the medical imaging study data structure:

segmenting the medical image to identify anatomical structures present within the medical images;

selecting a region of interest (ROI) in the medical image; and

extracting topological features of the ROI, wherein the temporal trajectory is generated based on changes in the topological features of the ROI for temporally subsequent medical images.

3. The method of claim 2 , wherein selecting medical image data from the medical imaging study data structure based on the temporal trajectory comprises analyzing the temporal trajectory data structure to identify salient portions of a temporal trajectory and selecting medical image data for the salient portions of the temporal trajectory.

4. The method of claim 2 , wherein the ROI is selected based on a purpose of a medical imaging study summary specified in a user request input to the medical imaging study summary engine, and a correlation of the purpose of the medical imaging study summary with anatomical structures in the medical image data of the medical imaging study data structure.

5. The method of claim 4 , wherein selecting the ROI further comprises comparing the medical image data of the medical imaging study data structure to previously labeled medical images corresponding to the purpose of the medical imaging study summary.

6. The method of claim 4 , wherein the topological features of the ROI that are extracted are topological features specific to the particular purpose of the medical imaging study summary, and wherein the topological features are measurements of anatomical structures present within the region of interest.

7. The method of claim 2 , wherein selecting the ROI comprises performing a multi-thresholding operation on an intensity histogram of the medical image data in the medical imaging study data structure and performing color connected component grouping on image characteristics of the medical image data to identify anatomical structures having similar characteristics.

8. The method of claim 2 , wherein extracting topological features of the ROI comprises:

extracting, from the ROI the first moment or area, second moments describing orientation, and a number of regions adjacent to another region;

generating a histogram of a range of values for each of the extracted topological features;

identifying peaks in the histogram to generate a multi-dimensional topological feature vector;

computing the multi-dimensional topological feature vector for each successive medical image in the medical image data of the medical imaging study data structure to generate a time-varying multi-dimensional topological feature vector curve;

performing a scale-space operation on the time varying multi-dimensional topological feature vector curve; and

identifying peaks in the time varying multi-dimensional topological feature vector curve that are above an optimal level identified by the scale-space operation, wherein the optimal level specifies a level of smoothing and level of approximation error.

9. The method of claim 1 , wherein selecting, by the medical imaging study summary engine, medical image data corresponding to selected medical images from the medical imaging study data structure based on a temporal trajectory data structure comprises selecting medical images from the medical imaging study data structure in which structures in the ROI of the selected medical images have characteristics that change above a predetermined level in a time series of medical images of the temporal trajectory data structure.

10. A computer program product comprising a non-transitory computer readable medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to implement a medical imaging study summary engine that operates to:

receive a medical imaging study data structure having data representing a plurality of medical images of a patient;

generate a temporal trajectory data structure of at least a subset of the medical images in the plurality of medical images, wherein the temporal trajectory data structure specifies topological changes in temporally subsequent medical images in the plurality of medical images;

select medical image data corresponding to selected medical images from the medical imaging study data structure based on the temporal trajectory data structure; and

output the selected medical image data via a medical imaging study user interface at least by generating a storyboard representation of the medical imaging study based on the selected medical images.

11. The computer program product of claim 10 , wherein the computer readable program further causes the medical imaging study summary engine to generate the temporal trajectory data structure at least by, for each medical image in the medical imaging study data structure:

segmenting the medical image to identify anatomical structures present within the medical images;

selecting a region of interest (ROI) in the medical image; and

extracting topological features of the ROI, wherein the temporal trajectory is generated based on changes in the topological features of the ROI for temporally subsequent medical images.

12. The computer program product of claim 11 , wherein the computer readable program further causes the medical imaging study summary engine to select medical image data from the medical imaging study data structure based on the temporal trajectory at least by analyzing the temporal trajectory data structure to identify salient portions of a temporal trajectory and selecting medical image data for the salient portions of the temporal trajectory.

13. The computer program product of claim 11 , wherein the ROI is selected based on a purpose of a medical imaging study summary specified in a user request input to the medical imaging study summary engine, and a correlation of the purpose of the medical imaging study summary with anatomical structures in the medical image data of the medical imaging study data structure.

14. The computer program product of claim 13 , wherein the computer readable program further causes the medical imaging study summary engine to select the ROI further at least by comparing the medical image data of the medical imaging study data structure to previously labeled medical images corresponding to the purpose of the medical imaging study summary.

15. The computer program product of claim 13 , wherein the topological features of the ROI that are extracted are topological features specific to the particular purpose of the medical imaging study summary, and wherein the topological features are measurements of anatomical structures present within the region of interest.

16. The computer program product of claim 11 , wherein the computer readable program further causes the medical imaging study summary engine to select the ROI at least by performing a multi-thresholding operation on an intensity histogram of the medical image data in the medical imaging study data structure and performing color connected component grouping on image characteristics of the medical image data to identify anatomical structures having similar characteristics.

17. The computer program product of claim 11 , wherein the computer readable program further causes the medical imaging study summary engine to extract topological features of the ROI at least by:

extracting, from the ROI the first moment or area, second moments describing orientation, and a number of regions adjacent to another region;

generating a histogram of a range of values for each of the extracted topological features;

identifying peaks in the histogram to generate a multi-dimensional topological feature vector;

computing the multi-dimensional topological feature vector for each successive medical image in the medical image data of the medical imaging study data structure to generate a time-varying multi-dimensional topological feature vector curve;

performing a scale-space operation on the time varying multi-dimensional topological feature vector curve; and

identifying peaks in the time varying multi-dimensional topological feature vector curve that are above an optimal level identified by the scale-space operation, wherein the optimal level specifies a level of smoothing and level of approximation error.

18. The computer program product of claim 10 , wherein the computer readable program further causes the medical imaging study summary engine to select medical image data corresponding to selected medical images from the medical imaging study data structure based on a temporal trajectory data structure at least by selecting medical images from the medical imaging study data structure in which structures in the ROI of the selected medical images have characteristics that change above a predetermined level in a time series of medical images of the temporal trajectory data structure.

19. An apparatus comprising:

a processor; and

a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to implement a medical imaging study summary engine that operates to:

receive a medical imaging study data structure having data representing a plurality of medical images of a patient;

generate a temporal trajectory data structure of at least a subset of the medical images in the plurality of medical images, wherein the temporal trajectory data structure specifies topological changes in temporally subsequent medical images in the plurality of medical images;

select medical image data corresponding to selected medical images from the medical imaging study data structure based on the temporal trajectory data structure; and

output the selected medical image data via a medical imaging study user interface at least by generating a storyboard representation of the medical imaging study based on the selected medical images.