Development of motion image prediction method using principal componentanalysis
Development of motion image prediction method using principal componentanalysis
著者:
Ritu Bhusal Chhatkuli Kazuyuki Demachi Masaki Kawai Hiroshi Sakakibara Kazuma Kamiaka
発刊日:
公開日:
image predictionlung tumorradiation therapytumor tracking
概要
Respiratory motion can induce the limit in the accuracy of area irradiated during lung cancer radiation therapy. Many methods have been introduced to minimize the impact of healthy tissue irradiation due to the lung tumor mot ion. The purpose of this research is to develop an algorithm for the improvement of image guided radiation therapy by the prediction of motion images. We predict the motion images by using principal component analysis (PCA) a nd multi-channel singular spectral analysis (MSSA) method. The images/movies were successfully predicted and verified using the developed algorithm. With the proposed prediction method it is possible to forecast the tumor ima ges over the next breathing period. The implementation of this method in real time is believed to be significant for higher level of tumor tracking including the detection of sudden abdominal changes during radiation therapy.
