Premium
SU‐FF‐J‐39: A Comparison of Two Methods Extracting Respiratory Information From Ultrasound Video
Author(s) -
Hu G,
Yao J,
Quan H,
Chen W
Publication year - 2007
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.2760544
Subject(s) - centroid , standard deviation , computer science , computer vision , ultrasound , artificial intelligence , biomedical engineering , nuclear medicine , medicine , mathematics , radiology , statistics
Purpose: To acquire respiratory information used for interventional radiotherapy in thorax and upper abdomen tumours and to compare two methods extracting respiratory information from ultrasound video. Method and Materials: Two test phantoms immersing in water tank were used to mimic moving tumours in intracavity. Phantoms were filled with a variety of wires and plastics to acquire whole images. One is sphere with 4.3 cm and the other is irregular object with maximum distance of 3.1cm and with minimum distance of 2.7cm. Motion range of two phantoms was on order of 3cm in 2D (radial section of ultrasound beam) and motion fashion was also examined in irregular motion as well as period varying from about 2s to 6s, for periodic. A comparison of two methods, correlation coefficient (CC) and quasi geometric centroid (QC) was implemented by metrics, namely, period, phase difference and relative standard deviation of period for period motion. For irregular, only phase difference is examined. Results: Results of comparison show that two methods are almost equal in these metrics, though small differences are exist due to some limited factors, such as imaging quality, selection of reference frame. QC is also appropriate for acquiring respiratory information. Conclusion: Acquisition of respiratory information from extraction video information is a superior method to the other methods which is noninvasive and reliable. Two methods used in the experiment demonstrate similar results, which indicate that QC is also competent to acquire respiratory information in various clinical circumstances. Due to the inherent performance of QC, more information can be obtained than that of correlation coefficient and thereby QC is possibly appropriate for accurate radiotherapy.