Abstract
Due to the risk of radiation from computed tomography (CT) scanning on the human body, the number of CT scans that can be performed on an individual each year is limited. However, CT images play a very important role in medical diagnosis. Therefore, this study proposes a method of generating synthetic CT to solve this problem. Considering that magnetic resonance imaging (MRI) is not harmful to the human body, there is no limit on the number of scans that can be performed with this procedure. In this paper, an image segmentation method is used to segment an MRI, and each segment is given a corresponding Hounsfield Unit (HU) value to finally generate a synthetic CT image. Since the image segmentation performance directly affects the generated synthetic CT image, this paper introduces a multitask learning strategy into a maximum entropy clustering (MEC) algorithm. A multitask maximum entropy clustering (MT-MEC) algorithm is proposed, which is used to effectively segment the MRI of the brain. The algorithm can use knowledge from multiple tasks to improve the learning ability of all tasks, and the MEC algorithm can effectively avoid interference from noise. The experimental results show that the proposed MT-MEC algorithm has good image segmentation performance, which results in reliable performance of the final synthetic CT image.
Highlights
It is well known that computed tomography (CT) scanning poses a radiation hazard to the human body
The experimental results show that the MT-maximum entropy clustering (MEC) algorithm is robust to noise, and the introduction of a multitask learning strategy takes into account the association between tasks, which improves the clustering performance of the algorithm and greatly improves the segmentation accuracy of the image
This study proposes a method of automatically generating a synthetic CT image
Summary
It is well known that CT scanning poses a radiation hazard to the human body. The number of CT scans an individual can receive each year is limited. The experimental results show that the MT-MEC algorithm is robust to noise, and the introduction of a multitask learning strategy takes into account the association between tasks, which improves the clustering performance of the algorithm and greatly improves the segmentation accuracy of the image. (3) Multitask maximum entropy clustering is applied to the brain MRI segmentation to improve the segmentation performance of the image and enhance the effect of the synthetic CT. To reduce the class center of each task as much as possible and obtain the optimal clustering effect of the multitask data, this paper proposes the following central reduction strategy: vj,k is very close to public op, and delete vj,k from private Vk ; vj,k is far away from public op, and take no operation; if rjp,k ≥. A detailed introduction to and description of this dataset can be found in our previous studies [50]–[52]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
