Children are usually examined using ultrasound to protect them from radiation exposure. Researchers at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA are now working on a system to automatically document the spatial recording position of ultrasound scans relative to the patient's body.
Several million ultrasound examinations are performed in Germany every year. As the human body is particularly sensitive to X-rays during growth, ultrasound is often used in pediatrics instead.
Until now, however, documenting ultrasound images has been a time-consuming process. For example, if doctors discover a cyst or tumor in a child's body, they first have to measure it and manually record its position in a 2D pictogram. This requires that the position of the ultrasound probe on the patient's body be transferred to the pictogram on the ultrasound system screen. Not only does this manual documentation take up roughly a quarter of treatment time in hectic everyday hospital practice; it is also potentially inaccurate.
In the SonoMap project, researchers at Fraunhofer IPA are now working on a significant improvement. “Using a 3D camera, we can not only automatically detect the recording position of the ultrasound probe and visualize it in three dimensions; we can also record the angle from which the image was taken,” explains Fraunhofer IPA research scientist Oliver Gölz. The side and angle from which an object is scanned makes a visible difference. “The additional tilt and angle information in the 3D image could help to locate tumors or cysts more quickly, making follow-up examinations more efficient.”
After a 3D camera, which is being used for the first time in this process, has detected the ultrasound probe and measured the surface of the body, AI-based image processing algorithms determine the position of the ultrasound probe and display it relative to the abstracted body surface. The system uses this information to generate an automatic 3D visualization that can be rotated and saved, enabling the patient's body to be viewed from different perspectives.
“Using our new system, doctors only have to save the ultrasound image, and the spatial documentation is generated automatically,” says Gölz. “This makes the process potentially faster and more accurate, leaving more time for patients.”
There are currently no comparable approaches on the ultrasound market. A demonstrator of the new technology is already complete. This will be used in a clinical study that Gölz and his team have applied for.
After completing the study, the researchers want to work with industry partners to equip ultrasound systems with the new function. They are currently ensuring that the 3D camera only takes secure images, ensuring patient data protection.
Gölz is delighted with success to date: “I am thrilled that we directly address problems from everyday hospital practice and relieve the burden on doctors.”
Fraunhofer-Gesellschaft