A 0–1 control mechanism with partial information for periodic tumour motion classification

In this paper, online treatment control mechanism for a stream of image classification is proposed. The developed two-layer mechanism controls the treatment window through a series of images acquisition. The paper contributes two findings for designing tumour treatment processes. Firstly, a theoretical foundation is established with numerical support in selecting an adaptive region for computation (RFC) with updated breathing pattern which is aimed to minimize the classification error. Secondly, the simulated results show that regardless image quality the proposed 0-1 treatment control mechanism could achieve comparable false positive rate ( <; 3.0%) and false negative rate ( <; 8.0%) to the best literature results which is based on fluoroscopic X-ray technology by classifying the tumour either inside or outside the treatment window.