Conclusions and Contributions - Prediction and Classification of Respiratory Motion - page 131

the feature selection metrics that are composed of a variety of breathing
features.
• The proposed CNN can outperform RNN with respect to the prediction accuracy
with an improvement of 50 %.
• CNN works for any of the five classes; thus, there are no particular differences
of error among the five classes because the criterion of feature sections in CNN
is designed to minimize the error.
• CNN does not directly address the criterion of overshoot regarding the class
selection among multiple patients; therefore, the larger size of patients may have
relatively large overshoot in the particular class.
The following conclusions can be made from the results obtained from Chap. 6 :
7.1.3 Irregular Breathing Classification from Multiple
Patient Datasets
• Irregular breathing patterns can be detected using neural networks, where the
reconstruction error can be used to build the distribution model for each
breathing class.
• The classifier using neural networks can provide clinical merit for the statisti-
cally quantitative modeling of irregular breathing motion based on a regular
ratio representing how many regular/irregular patterns exist within an obser-
vation period.
• The breathing data can be categorized into several classes based on the extracted
feature metrics derived from the breathing data of multiple patients.
• The breathing cycles are distributed with a minimum of 2.9 s/cycle, a maximum
of 5.94 s/cycle, and the average breathing cycle of 3.91 s/cycle.
• The breathing pattern for each patient can be classified into regular/irregular
breathing using the regular ratio, even though the breathing data are mixed up
with the regular and irregular breathing patterns in the given samples.
• The true positive rate (TPR) for the proposed classifier is 97.83 % when the
false positive rate (FPR) is 50 %.
• The breathing cycles of any given patient with a length of at least 900 s can be
classified reliably enough to adjust the safety margin prior to therapy in the
proposed classification.
7.2 Contributions
This study has three main contributions on the prediction of respiratory motion in
radiation therapy.