Objective To address the challenges of accurately capturing critical features in small-sample diabetic retinopathy(DR)recognition models in real-world applications,and the overly smooth distribution of true and false feature coefficients,we propose an enhanced small-sample DR recognition method based on an improved capsule network.Methods Firstly,the method enhances image feature representation by removing redundant boundary information and employing discrete wavelet transform based on the Haar wavelet function,thereby high-lighting critical pathological features.Secondly,the convolutional layer of the capsule network is optimized through a multi-branch architecture to extract multi-scale features from retinal images,while incorporating a convolutional block attention module that is subsequently fed into the capsule layer.Finally,the sigmoid function replaces the soft-max function in dynamic routing,thereby improving the model's robustness.Result The enhanced neural network model achieved an accuracy of 98.62%on the Kaggle public dataset following a rigorous selection and preprocessing procedure.Conclusion The enhanced capsule network demonstrated superior precision in identifying diabetic reti-nopathy within small sample sizes compared to other state-of-the-art algorithms currently available.

Objective To address the challenges of accurately capturing critical features in small-sample diabetic retinopathy(DR)recognition models in real-world applications,and the overly smooth distribution of true and false feature coefficients,we propose an enhanced small-sample DR recognition method based on an improved capsule network.Methods Firstly,the method enhances image feature representation by removing redundant boundary information and employing discrete wavelet transform based on the Haar wavelet function,thereby high-lighting critical pathological features.Secondly,the convolutional layer of the capsule network is optimized through a multi-branch architecture to extract multi-scale features from retinal images,while incorporating a convolutional block attention module that is subsequently fed into the capsule layer.Finally,the sigmoid function replaces the soft-max function in dynamic routing,thereby improving the model's robustness.Result The enhanced neural network model achieved an accuracy of 98.62%on the Kaggle public dataset following a rigorous selection and preprocessing procedure.Conclusion The enhanced capsule network demonstrated superior precision in identifying diabetic reti-nopathy within small sample sizes compared to other state-of-the-art algorithms currently available.

The aim of this study is to synthesize the ordered mesoporous silica (OMS) as drug carrier to improve release property of insoluble drug and investigate the dissolution profile of insoluble drug from the porous carrier. The OMS was obtained by using cetyltrimethyl ammonium bromide as the template and resveratrol was selected as the model drug. The resveratrol-loaded OMS (Res-OMS) were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), N2 adsorption-desorption, X-ray diffraction (XRD) and FT-IR spectroscopy. In vitro drug release behavior was also investigated. It was found that the synthesized OMS showed a large surface area, a narrow pore size distribution and an important mesoporosity associated to hexagonally organized channels. Compared with physical mixture and crystalline powder, resveratrol was in amorphous or molecular form after loading into OMS. The release rate ofresveratrol from drug-loaded OMS was significantly increased suggesting the great potential application of OMS for the formulation of poorly soluble drugs.

Objective To study the intervention effect of tanshinone on electrophysiological abnormality of hypertrophic cardicoyte in order to illuminate the underlying mechanism of tanshinone in preventing the arrhythmia induced by myocardial hypertrophy. Method Twenty-week-rid SD rats (200～250 g) were divided into 4 groups (8 in each group) randomly. Of 4 groups, rats of three groups were operated on by a procedure of 'one kidney one clamp' to make renal artery constriction. The rest group served as sham operation group (control group). When the blood pressure increased,rats of operation groups were divided into tanshinone group, captopril group and hyper-trophic group. The effects of tanshinoe and captopril were observed and compared on the action potential duration (APD),L-type calcium current (ICa, L) and transient outward potassium current (Ito) density in cellular membrane of hypertrophic myocardium by using patch clamp and intra-cellular calcium survey technique. Results The blood pressure in operation groups was obviously higher than that in sham-operation group (P<0.01), but there was no difference between operation groups (P>0.05). The ratio of ventricle weight to body weight (VW/BW) was much higher in hypertrophic group than in control group (P<0.01), and it significantly decreased after interven-tion with tanshinone or captopril (P<0.01). Compared with hypertrophic group, tanshinone markedly shortened the prolongation of action potential duration (P<0.01), decreased membrane capacity and peak amplitude of ICa,L(P<0.01), but had no effect on the density of ICa,L. Tanshinone also significantly increased Ito current density and peak amplitude, which were completely different from hypertrophic group (P<0.05). There were similar results foundin captopril intervention. Conclusions Tanshinone could reduce calcium influx and resume the activity of ho ion channels, and thus shorten the first phase and the plateau phase of repolarization and decrease the prolongation of APD in hypertrophic cadiocyte. So tanshinone can prevent the onset of arrhythmia attributed to the myocardial hypertrophy.