Prediction method for healthy radius of blood vessel path, prediction method for candidate stenosis of blood vessel path, and blood vessel stenosis degree prediction device
A computer-implemented method for predicting a blood vessel stenosis is disclosed. The method may include extracting a blood vessel path and its centerline based on the image of the blood vessel. The method may further include determining a candidate stenosis for the blood vessel path and identifying image blocks along the centerline of the blood vessel path within a range of candidate stenosis for the blood vessel path determined based on the candidate stenosis. The method may also include determining a degree of stenosis for the blood vessel path by applying a trained learning network comprising a convolutional neural network and a recurrent neural network on the image blocks within the range of candidate stenosis.
1. A computer-implemented method for predicting a blood vessel stenosis, the method comprising:
extracting a blood vessel path and its centerline based on an image of a blood vessel;
determining a candidate stenosis for the blood vessel path;
identifying image blocks along the centerline of the blood vessel path within a range of candidate stenosis for the blood vessel path determined based on the candidate stenosis; and
determining a degree of stenosis for the blood vessel path by applying a trained learning network comprising a convolutional neural network and a recurrent neural network on the image blocks within the range of candidate stenosis,
wherein the convolutional neural network and the recurrent neural network are sequentially applied, wherein the convolutional neural network is applied on the image blocks along the centerline of the blood vessel path to generate vectors for the respective image blocks, wherein the recurrent neural network is applied on the vectors.
2. The computer-implemented method according to claim 1 , further comprising:
predicting a healthy radius of the blood vessel path by performing a quadratic regression on blood vessel radius peaks of the blood vessel path; and
determining the candidate stenosis for the blood vessel path based on the healthy radius of the blood vessel path.
3. The computer-implemented method according to claim 2 , wherein predicting the healthy radius of the blood vessel path further comprises:
predicting a reference healthy radius of the blood vessel path by performing a linear regression on the blood vessel radius peaks; and
replacing the blood vessel radius peaks in the blood vessel path that are lower than the reference healthy radius with the reference healthy radius.
4. The computer-implemented method according to claim 1 , wherein the range of candidate stenosis for the blood vessel path has the candidate stenosis at its center.
5. The computer-implemented method according to claim 1 , wherein the convolutional neural network and the recurrent neural network are jointly trained.
6. The computer-implemented method according to claim 1 , wherein the convolutional neural network comprises a plurality of convolutional neural network units each configured to process one of the image blocks along the centerline of the blood vessel path.
7. The computer-implemented method according to claim 1 , wherein the recurrent neural network comprises a forward recurrent neural network and a reverse recurrent neural network configured to collectively learn a spatial correlation among the image blocks along the centerline of the blood vessel path.
8. The computer-implemented method according to claim 7 , wherein the forward recurrent neural network and the reverse recurrent neural network are separately trained.
9. The computer-implemented method according to claim 1 , further comprising outputting at least one of the candidate stenosis, the range of candidate stenosis, and the degree of stenosis of the blood vessel path.
10. A blood vessel stenosis degree prediction system, comprising:
an interface configured to receive an image of a blood vessel;
a processor configured to:
extract a blood vessel path and its centerline based on the image of the blood vessel;
determine a candidate stenosis for the blood vessel path;
identify image blocks along the centerline of the blood vessel path within a range of candidate stenosis for the blood vessel path determined based on the candidate stenosis; and
determine a degree of stenosis for the blood vessel path by applying a trained learning network comprising a convolutional neural network and a recurrent neural network on the image blocks within the range of candidate stenosis,
wherein the convolutional neural network and the recurrent neural network are sequentially applied, wherein the convolutional neural network is applied on the image blocks along the centerline of the blood vessel path to generate vectors for the respective image blocks, wherein the recurrent neural network is applied on the vectors.
11. The blood vessel stenosis degree prediction system according to claim 10 , wherein the processor is further configured to:
predict a healthy radius of the blood vessel path by performing a quadratic regression on blood vessel radius peaks of the blood vessel path; and
determine the candidate stenosis for the blood vessel path based on the healthy radius of the blood vessel path.
12. The blood vessel stenosis degree prediction system according to claim 10 , wherein to predict the healthy radius of the blood vessel path, the processor is further configured to:
predict a reference healthy radius of the blood vessel path by performing a linear regression on the blood vessel radius peaks; and
replace the blood vessel radius peaks in the blood vessel path that are lower than the reference healthy radius with the reference healthy radius.
13. The blood vessel stenosis degree prediction system according to claim 10 , wherein the range of candidate stenosis for the blood vessel path has the candidate stenosis at its center.
14. The blood vessel stenosis degree prediction system according to claim 10 , wherein the convolutional neural network and the recurrent neural network are jointly trained.
15. The blood vessel stenosis degree prediction system according to claim 10 , wherein the recurrent neural network comprises a forward recurrent neural network and a reverse recurrent neural network configured to collectively learn a spatial correlation among the image blocks along the centerline of the blood vessel path.
16. The blood vessel stenosis degree prediction system according to claim 15 , wherein the forward recurrent neural network and the reverse recurrent neural network are separately trained.
17. The blood vessel stenosis degree prediction system according to claim 10 , further comprising a display configured to output at least one of the candidate stenosis, the range of candidate stenosis, and the degree of stenosis of the blood vessel path.
18. The blood vessel stenosis degree prediction system according to claim 10 , wherein the convolutional neural network comprises a plurality of convolutional neural network units each configured to process one of the image blocks along the centerline of the blood vessel path.
19. A non-transitory storage medium having stored thereon computer executable instructions that, when executed by a processor, implement a method for predicting a blood vessel stenosis, the method comprising:
extracting a blood vessel path and its centerline based on an image of a blood vessel;
determining a candidate stenosis for the blood vessel path;
identifying image blocks along the centerline of the blood vessel path within a range of candidate stenosis for the blood vessel path determined based on the candidate stenosis; and
determining a degree of stenosis for the blood vessel path by applying a trained learning network comprising a convolutional neural network and a recurrent neural network on the image blocks within the range of candidate stenosis,
wherein the convolutional neural network and the recurrent neural network are sequentially applied, wherein the convolutional neural network is applied on the image blocks along the centerline of the blood vessel path to generate vectors for the respective image blocks, wherein the recurrent neural network is applied on the vectors.
20. The non-transitory storage medium according to claim 19 , wherein the method further comprises outputting at least one of the candidate stenosis, the range of candidate stenosis, and the degree of stenosis of the blood vessel path.