ABSTRACT
PURPOSE: In high-intensity focused ultrasound (HIFU) treatment of the kidney and liver, tracking the organs is essential because respiratory motions make continuous cauterization of the affected area difficult and may cause damage to other parts of the body. In this study, we propose a tracking system for rotational scanning, and propose and evaluate a method for estimating the angles of organs in ultrasound images. METHODS: We proposed AEMA, AEMAD, and AEMAD++ as methods for estimating the angles of organs in ultrasound images, using RUDS and a phantom to acquire 90-degree images of a kidney from the long-axis image to the short-axis image as a data set. Six datasets were used, with five for preliminary preparation and one for testing, while the initial position was shifted by 2 mm in the contralateral axis direction. The test data set was evaluated by estimating the angle using each method. RESULTS: The accuracy and processing speed of angle estimation for AEMA, AEMAD, and AEMAD++ were 23.8% and 0.33 FPS for AEMAD, 32.0% and 0.56 FPS for AEMAD, and 29.5% and 3.20 FPS for AEMAD++, with tolerance of ± 2.5 degrees. AEMAD++ offered the best speed and accuracy. CONCLUSION: In the phantom experiment, AEMAD++ showed the effectiveness of tracking the long-axis image of the kidney in rotational scanning. In the future, we will add either the area of surrounding organs or the internal structure of the kidney as a new feature to validate the results.
ABSTRACT
PURPOSE: An inevitable feature of ultrasound-based diagnoses is that the quality of the ultrasound images produced depends directly on the skill of the physician operating the probe. This is because physicians have to constantly adjust the probe position to obtain a cross section of the target organ, which is constantly shifting due to patient respiratory motions. Therefore, we developed an ultrasound diagnostic robot that works in cooperation with a visual servo system based on deep learning that will help alleviate the burdens imposed on physicians. METHODS: Our newly developed robotic ultrasound diagnostic system consists of three robots: an organ tracking robot (OTR), a robotic bed, and a robotic supporting arm. Additionally, we used different image processing methods (YOLOv5s and BiSeNet V2) to detect the target kidney location, as well as to evaluate the appropriateness of the obtained ultrasound images (ResNet 50). Ultimately, the image processing results are transmitted to the OTR for use as motion commands. RESULTS: In our experiments, the highest effective tracking rate (0.749) was obtained by YOLOv5s with Kalman filtering, while the effective tracking rate was improved by about 37% in comparison with cases without such filtering. Additionally, the appropriateness probability of the ultrasound images obtained during the tracking process was also the highest and most stable. The second highest tracking efficiency value (0.694) was obtained by BiSeNet V2 with Kalman filtering and was a 75% improvement over the case without such filtering. CONCLUSION: While the most efficient tracking achieved is based on the combination of YOLOv5s and Kalman filtering, the combination of BiSeNet V2 and Kalman filtering was capable of detecting the kidney center of gravity closer to the kidney's actual motion state. Furthermore, this model could also measure the cross-sectional area, maximum diameter, and other detailed information of the target kidney, which meant it is more practical for use in actual diagnoses.
