Method and system for the automated determination of examination results in an image sequence
One or more example embodiments of the present invention relates to a method for the automated determination of examination results in an image sequence from multiple chronologically consecutive frames, the method comprising determining diagnostic candidates in the form of contiguous image regions in the individual frames for a predefined diagnostic finding; and for a number of the diagnostic candidates, determining which candidate image regions in other frames correspond to the particular diagnostic candidate, determining whether the candidate image regions of the particular diagnostic candidate in the other frames overlap with other diagnostic candidates, generating a graph containing the determined diagnostic candidates of the frames as nodes and the determined overlaps as edges, and generating communities from nodes connected via edges.
1. A method for an automated determination of examination results in an image sequence from multiple chronologically consecutive frames, the method comprising:
determining diagnostic candidates as contiguous image regions in individual frames for a predefined diagnostic finding; and
for a number of the diagnostic candidates,
determining which candidate image regions in other frames correspond to a particular diagnostic candidate,
determining whether the candidate image regions of the particular diagnostic candidate in the other frames overlap with other diagnostic candidates,
generating a graph containing the determined diagnostic candidates of the frames as nodes and the determined overlaps as edges, and
generating communities from nodes connected via edges.
2. The method of claim 1 , further comprising:
outputting the communities or using the communities for further examinations.
3. The method as claimed in claim 2 , wherein the further examinations include an X-ray angiography examination.
4. The method of claim 1 , wherein
the diagnostic candidates are representatives of medical diagnostic findings.
5. The method of claim 4 , wherein the medical diagnostic findings are in the field of angiography, wherein at least one diagnostic candidate represents the finding of a vascular irregularity.
6. The method of claim 1 , wherein the determining the diagnostic candidates comprises:
generating a feature map in an examined frame, the feature maps including regions that are marked with at least one probability value for a corresponding diagnostic candidate of the diagnostic candidates;
forming contiguous image regions based on the feature map; and
output of the regions as the diagnostic candidates.
7. The method of claim 6 , wherein the forming includes,
comparing the probability values of the feature map to a limit value, the contiguous image regions including the image regions in which the probability values exceed the limit value.
8. The method of claim 6 , wherein the generating the feature map uses an image-to-image convolution network, the image-to-image convolution network being trained with real frames with manually marked diagnostic candidates.
9. The method of claim 1 , wherein the determining the corresponding candidate image regions in other frames includes,
assessing a deformation field including modeling a movement between a first frame and a second frame; and
registering of an image region of the diagnostic candidate of the first frame to a candidate image region of the second frame using the deformation field.
10. The method of claim 1 , further comprising:
weighting the edges, the weighting being based on at least one of
a degree of overlap,
a chronological interval between the frames,
a similarity,
a probability, or
an uncertainty value.
11. The method of claim 1 , wherein the generating generates the communities by
trimming at least one of the edges or the nodes based on predefined criteria, or using community algorithms.
12. The method of claim 1 , wherein the generating the communities includes,
assigning multiple diagnostic findings in the same frame to a same community.
13. The method of claim 1 , further comprising:
evaluating a structure of at least one of the communities, the evaluating including selecting one frame in the image sequence as a key frame, wherein the key frame is selected for an identified diagnostic finding by determining a node that is has maximum support within the community.
14. The method of claim 1 , wherein the method is performed for a plurality of image sequences from multiple chronologically consecutive frames acquired from different acquisition positions or during different examinations, and the diagnostic candidates from the image sequences are combined in the graph.
15. A non-transitory computer program product comprising commands which when the program is executed by a computer cause said computer to execute the method of claim 1 .
16. A non-transitory computer-readable medium comprising commands which when executed by a computer cause said computer to execute the method of claim 1 .
17. A system for an automated determination of examination results in an image sequence from multiple chronologically consecutive frames, the system comprising:
a diagnostic study unit configured to determine diagnostic candidates as contiguous image regions in individual frames for a predefined diagnosis;
a propagation unit configured to determine for a diagnostic candidate in a frame as to which candidate image regions in other frames correspond to the diagnostic candidate;
a relation unit configured to determinate overlaps, the overlaps indicating whether the candidate image regions of the diagnostic candidate in the other frames overlap with other diagnostic candidates;
a graph unit configured to generate a graph containing the determined diagnostic candidates in the frames as nodes and the determined overlaps as edges;
a community unit configured to generate communities from the nodes connected via edges; and
one of,
an output unit configured to output the communities, or
a diagnostic study unit configured to further examine the communities.
18. The system of claim 17 , wherein the system is configured to process a plurality of image sequences from multiple chronologically consecutive frames acquired from different acquisition positions or during different examinations, and the diagnostic candidates from the image sequences are combined in the graph.
19. The system of claim 17 , further comprising:
a key frame unit configured to determine a key frame using properties of a structure of at least one community.
20. A medical technology system comprising the system of claim 17 .