In this paper, we propose Hformer, a novel supervised learning model for low-dose computer tomography (LDCT) denoising. Hformer combines the strengths of convolutional neural networks for local feature extraction and transformer models for global feature capture. The performance of Hformer was verified and evaluated based on the AAPM-Mayo Clinic LDCT Grand Challenge Dataset. Compared with the former representative state-of-the-art (SOTA) model designs under different architectures, Hformer achieved optimal metrics without requiring a large number of learning parameters, with metrics of 33.4405 PSNR, 8.6956 RMSE, and 0.9163 SSIM. The experiments demonstrated designed Hformer is a SOTA model for noise suppression, structure preservation, and lesion detection.
Carbon ions, commonly referred to as particle therapy, have become increasingly popular in the last decade. Accurately predicting the range of ions in tissues is important for the precise delivery of doses in heavy-ion radiotherapy. Range uncertainty is currently the largest contributor to dose uncertainty in normal tissues, leading to the use of safety margins in treatment planning. One potential method is the direct relative stopping measurement (RSP) with ions. Heavy-ion CT (Hi′CT), a compact segmented full digital tomography detector using monolithic active pixel sensors, was designed and evaluated using a 430 MeV/u high-energy carbon ion pencil beam in Geant4. The precise position of the individual carbon ion track can be recorded and reconstructed using a 30 μm × 30 μm small pixel pitch size. Two types of customized image reconstruction algorithms were developed, and their performances were evaluated using three different modules of CATPHAN 600-series phantoms. The RSP measurement accuracy of the tracking algorithm for different types of materials in the CTP404 module was less than 1%. In terms of spatial resolution, the tracking algorithm could achieve a 20% modulation transfer function normalization value of CTP528 imaging results at 5 lp/cm, which is significantly better than that of the fast imaging algorithm (3 lp/cm). The density resolution obtained using the tracking algorithm of the customized CTP515 was approximately 10.5%. In conclusion, a compact digital Hi'CT system was designed, and its nominal performance was evaluated in a simulation. The RSP resolution and image quality provide potential feasibility for scanning most parts of an adult body or pediatric patient, particularly for head and neck tumor treatment.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.