Machine learning super-resolution of laboratory CT images in all-solid-state batteries using synchrotron radiation CT as training data

High-performance all-solid-state lithium-ion batteries require observation,control,and optimization of the electrode structure.X-ray computational tomography(CT)is an effective nondestructive method for observing the electrode structure in three dimensions.However,the limited availability of synchrotron radiation CT,which offers high-resolution imaging with a high signal-to-noise ratio,makes it difficult to conduct experiments and restricts the use of X-ray CT in battery development.Conversely,laboratory CT systems are widely available,but they use X-rays emitted from a metal target,resulting in lower image quality and resolution compared with synchrotron radiation CT.This study explores a method for achieving comparable resolution in laboratory CT images of all-solid-state batteries to that of synchrotron radiation CT.Our method involves using the synchrotron radiation CT images as training data for machine learning super-resolution.The results demonstrate that,by employing an appropriate machine learning algorithm and activation function,along with a sufficiently deep network,the image quality of laboratory CT becomes equivalent to that of synchrotron radiation CT.

Early Imaging Features and Differential Diagnosis of Novel Coronavirus Pneumonia

Objective:To discuss the early imaging features of novel coronary pneumonia(NCP)and its differential diagnosis with common pneumonia for the clinical Provide relatively correct imaging diagnosis.Methods:A review of 10 cases of novel coronavirus pneumonia diagnosed in our hospital and surrounding counties was collected,and our hospital's 2019-2020 Common pneumonia such as influenza A and B virus pneumonia,lobar pneumonia and adenovirus confirmed by laboratory tests and abnormal chest radiographs Ten patients each with pneumonia,a total of 40 patients,were collected and their imaging features were analyzed.Results:In 10 patients with neo-coronary pneumonia,there were 30 lesions on chest CT,with typical characteristic lesions containing bronchograms within their Angiographic thickening;located in the subpleura,with grinding glass-like or combined solid changes,referred to as"extratubular halo",with multifocal and multifocal distribution.Morphology or unilobular large lamellar foci without lymph node enlargement and pleural effusion;10 cases of influenza A and B virus pneumonia in chest CT performance In nine patients,the lesions were distributed in the subpleural or along the perimeter of the bronchial vessels in the form of an analogous circular ground-glass shadow,some of which was a small piece of solid shadow,1 The case involved a single lobe of the lung and showed a large mixed ground glass image,and the CT in 10 cases of lobar pneumonia showed that they all had a single large solid lobe The variegated shadow or patchy cloud with blurred margins was triangularly altered with the tip pointing to the lung portal,and there were five cases of air-containing bronchograms.Adenoviral pneumonia is more common in infants and young children,and CT showed single or multiple grinded glass images in both lungs with patchy solid variegated shadows and lobar distribution.Conclusion:NCP and common pneumonia have certain imaging features that,in combination with laboratory tests and epidemiologic history,allow a preliminary diagnosis to be made.It has certain directions and help for clinical diagnosis.