Computer vision systems and methods for real-time localization of needles in ultrasound images
Computer vision systems and methods for real-time localization of a needle in an ultrasound video are provided. The method includes a step of receiving at a processor a plurality of frames from the ultrasound video. The method also includes the step of identifying by the processor one of the plurality of frames as a background image and a frame adjacent to the background image as a current image. The method further includes the step of performing by the processor a bitwise complement operation on the background image to generate a complement image. Moreover, the method includes the step of performing by the processor a pointwise logical AND operation on the complement image and the current image to identify the location of the needle.
1. A method for real-time localization of a medical device in a video, comprising the steps of:
receiving at a processor a plurality of frames from the video;
identifying by the processor one of the plurality of frames as a background image and a frame adjacent to the background image as a current image;
performing by the processor a bitwise complement operation on the background image to generate a complement image; and
performing by the processor a pointwise logical AND operation on the complement image and the current image to identify the location of the medical device.
2. The method of claim 1 , further comprising the step of enhancing the medical device using a Split Bregman process.
3. The method of claim 1 , wherein the medical device is a needle.
4. The method of claim 3 , wherein a tip of the needle is inserted in-plane.
5. The method of claim 3 , wherein a tip of the needle is inserted out-of-plane.
6. The method of claim 3 , wherein the needle's insertion angle is not known.
7. The method of claim 3 , wherein the needle is non-linear.
8. The method of claim 1 , further comprising the step of setting a threshold correlation coefficient for skipping a frame due to insignificant motion.
9. The method of claim 1 , wherein the video is an ultrasound video.
10. The method of claim 1 , further comprising the step of enhancing image data representing the medical device using a convolution neural network.
11. A system for real-time localization of a medical device in a video, comprising:
a processor; and
a memory in communication with the processor, the memory having computer instructions stored thereon which, when executed, cause the processor to perform the following steps:
receiving a plurality of frames from the video;
identifying one of the plurality of frames as a background image and a frame adjacent to the background image as a current image;
performing a bitwise complement operation on the background image to generate a complement image; and
performing a pointwise logical AND operation on the complement image and the current image to identify the location of the medical device.
12. The system of claim 11 , wherein the instructions further cause the processor to enhance the medical device using a Split Bregman process.
13. The system of claim 11 , wherein the medical device is a needle.
14. The system of claim 13 , wherein a tip of the needle is inserted in-plane.
15. The system of claim 13 , wherein a tip of the needle is inserted out-of-plane.
16. The system of claim 13 , wherein the needle's insertion angle is not known.
17. The system of claim 13 , wherein the needle is non-linear.
18. The system of claim 11 , wherein a threshold correlation coefficient is set for skipping a frame due to insignificant motion.
19. The system of claim 11 , wherein the video is an ultrasound video.
20. The system of claim 11 , wherein the instructions further cause the processor to enhance image data representing the medical device using a convolution neural network.
21. A non-transitory computer-readable medium having computer-readable instructions stored thereon which, when executed by a computer system, cause the computer system to perform the steps of:
receiving at a processor a plurality of frames from a video;
identifying by the processor one of the plurality of frames as a background image and a frame adjacent to the background image as a current image;
performing by the processor a bitwise complement operation on the background image to generate a complement image; and
performing by the processor a pointwise logical AND operation on the complement image and the current image to identify the location of a medical device.
22. The non-transitory computer readable medium of claim 21 , further comprising the step of enhancing the medical device using a Split Bregman process.
23. The non-transitory computer readable medium of claim 21 , wherein the medical device is a needle.
24. The non-transitory computer readable medium of claim 23 , wherein a tip of the needle is inserted in-plane.
25. The non-transitory computer readable medium of claim 23 , wherein a tip of the needle is inserted out-of-plane.
26. The non-transitory computer readable medium of claim 23 , wherein the needle's insertion angle is not known.
27. The non-transitory computer readable medium of claim 23 , wherein the needle is non-linear.
28. The non-transitory computer readable medium of claim 21 , further comprising the step of setting a threshold a correlation coefficient for skipping a frame due to insignificant motion.
29. The non-transitory computer readable medium of claim 21 , further comprising the step of enhancing image data representing the medical device using a convolution neural network.