Deep learning-based automatic classification of diabetic retinopathy (DR) helps to enhance the accuracy and efficiency of auxiliary diagnosis. This paper presents an improved residual network model for classifying DR into five different severity levels. First, the convolution in the first layer of the residual network was replaced with three smaller convolutions to reduce the computational load of the network. Second, to address the issue of inaccurate classification due to minimal differences between different severity levels, a mixed attention mechanism was introduced to make the model focus more on the crucial features of the lesions. Finally, to better extract the morphological features of the lesions in DR images, cross-layer fusion convolutions were used instead of the conventional residual structure. To validate the effectiveness of the improved model, it was applied to the Kaggle Blindness Detection competition dataset APTOS2019. The experimental results demonstrated that the proposed model achieved a classification accuracy of 97.75% and a Kappa value of 0.971 7 for the five DR severity levels. Compared to some existing models, this approach shows significant advantages in classification accuracy and performance.
Deformable image registration plays a crucial role in medical image analysis. Despite various advanced registration models having been proposed, achieving accurate and efficient deformable registration remains challenging. Leveraging the recent outstanding performance of Mamba in computer vision, we introduced a novel model called MCRDP-Net. MCRDP-Net adapted a dual-stream network architecture that combined Mamba blocks and convolutional blocks to simultaneously extract global and local information from fixed and moving images. In the decoding stage, we employed a pyramid network structure to obtain high-resolution deformation fields, achieving efficient and precise registration. The effectiveness of MCRDP-Net was validated on public brain registration datasets, OASIS and IXI. Experimental results demonstrated significant advantages of MCRDP-Net in medical image registration, with DSC, HD95, and ASD reaching 0.815, 8.123, and 0.521 on the OASIS dataset and 0.773, 7.786, and 0.871 on the IXI dataset. In summary, MCRDP-Net demonstrates superior performance in deformable image registration, proving its potential in medical image analysis. It effectively enhances the accuracy and efficiency of registration, providing strong support for subsequent medical research and applications.
Objective To study the characteristics of pulmonary impact injury under closure and open states of glottis. Methods One hundred and eight rabbits were randomly divided into two groups (54 each group). Open state of glottis group(open group): impact injuries with opened glottis; closure state of glottis group (closed group): impact injuries with closed glottis. Parameters were set up with various combinations of driven pressures and compress percentage and the model of rabbit blunt chest trauma were established. Pathological changes were examined and abbreviated injury scale (AIS), water containing and mortality were recorded. Results Two and four rabbits died in open group and closed group respectively under the condition of 30% for compress percentage and 8 250 mmHg for driven pressures. In most cases, AIS values of closed group were significantly higher than that of open group (Plt;0.05). AIS values were positively related to driven pressures and compress percentage (r=0.9313, 0.7847; Plt;0.01, 0.01). Quantities of contained water in lung of closed group were significantly higher than that of open group(t=2.28,Plt;0.01). Conclusion The severity of injury, the increased mortality and earlier occurrence of traumatic acute lung injury were the characteristics of pulmonary impact injury under the closure states of glottis.
The electroencephalogram (EEG) has proved to be a valuable tool in the study of comprehensive conditions whose effects are manifest in the electrical brain activity, and epilepsy is one of such conditions. In the study, multi-scale permutation entropy (MPE) was proposed to describe dynamical characteristics of EEG recordings from epilepsy and healthy subjects, then all the characteristic parameters were forwarded into a support vector machine (SVM) for classification. The classification accuracies of the MPE with SVM were evaluated by a series of experiments. It is indicated that the dynamical characteristics of EEG data with MPE could identify the differences among healthy, inter-ictal and ictal states, and there was a reduction of MPE of EEG from the healthy and inter-ictal state to the ictal state. Experimental results demonstrated that average classification accuracy was 100% by using the MPE as a feature to characterize the healthy and seizure, while 99.58% accuracy was obtained to distinguish the seizure-free and seizure EEG. In addition, the single-scale permutation entropy (PE) at scales 1-5 was put into the SVM for classification at the same time for comparative analysis. The simulation results demonstrated that the proposed method could be a very powerful algorithm for seizure prediction and could have much better performance than the methods based on single scale PE.
[Abstract]Automatic and accurate segmentation of lung parenchyma is essential for assisted diagnosis of lung cancer. In recent years, researchers in the field of deep learning have proposed a number of improved lung parenchyma segmentation methods based on U-Net. However, the existing segmentation methods ignore the complementary fusion of semantic information in the feature map between different layers and fail to distinguish the importance of different spaces and channels in the feature map. To solve this problem, this paper proposes the double scale parallel attention (DSPA) network (DSPA-Net) architecture, and introduces the DSPA module and the atrous spatial pyramid pooling (ASPP) module in the “encoder-decoder” structure. Among them, the DSPA module aggregates the semantic information of feature maps of different levels while obtaining accurate space and channel information of feature map with the help of cooperative attention (CA). The ASPP module uses multiple parallel convolution kernels with different void rates to obtain feature maps containing multi-scale information under different receptive fields. The two modules address multi-scale information processing in feature maps of different levels and in feature maps of the same level, respectively. We conducted experimental verification on the Kaggle competition dataset. The experimental results prove that the network architecture has obvious advantages compared with the current mainstream segmentation network. The values of dice similarity coefficient (DSC) and intersection on union (IoU) reached 0.972 ± 0.002 and 0.945 ± 0.004, respectively. This paper achieves automatic and accurate segmentation of lung parenchyma and provides a reference for the application of attentional mechanisms and multi-scale information in the field of lung parenchyma segmentation.
Telerehabilitation is a new rehabilitation technology, using internet to provide rehabilitation services for patients in remote areas or unaccessible to rehabilitation. Longshi Ability of Daily Life scale is fomulated based on Chinese living customs. The assessment content of the scale can clearly reflect the needs of the service object, and the assessment result can directly reflect the ability level of the assessment object. The scale has been put into use online on the mobile internet and amassed a certain amount of big data, which is of great significance to the adjustment of rehabilitation treatment, the continuity of nursing guidance, and the assurance of adequate social support and disability benefits for the disabled. In this paper, the application of Longshi Ability of Daily Life scale in telerehabilitation is described.
ObjectiveTo understand the application of the Braden pressure ulcer risk-factor assessment scale in the nursing staff, in order to provide reference for clinical nurses to standardize the use of Braden assessment scale and facilitate the hospital to develop training programs on pressure ulcer related knowledge. MethodsStratified cluster sampling method was applied in February 2015. Using the self-designed questionnaire of “Application of Braden pressure ulcer risk-factor assessment scale in the nursing staff ”, we conducted a survey on 198 clinical nurses, and the survey results were scrutinized. The difficulty level of using Braden assessment scale in the nurses was analyzed based on their different demographic characteristics. We also analyzed the items which were most difficult to judge for the nurses and nurses’ learning needs for knowledge on Braden assessment scale. ResultsA total of 168 (84.85%) nurses found it difficult in using Braden scale for the evaluation of pressure ulcer. The most difficult items to judge for the nurses were friction force, shear force and feeling. Nurses in departments with pressure ulcer as a common symptom of the patients could better use the Braden pressure ulcer risk-factor scale, compared with those in departments where pressure ulcer was uncommon (P＜ 0.05) . A total of 189 (95.46%) nurses thought it necessary to carry out a unified quantitative standard analysis of six risk factors in the Braden scale. Conclusions The poor mastery of the assessment standards for Braden scale in the nurses causes various degrees of difficulty in applying the scale, which can influence the accuracy of assessment. It is important to train the nurses on pressure ulcer risk factor assessment in order to raise the clinical assessment accuracy.
Objective To explore the correlation between risk factors in respiratory department patients and the occurrence of venous thromboembolism (VTE), and to evaluate the optimization of the Padua score for predicting VTE occurrence in hospitalized respiratory patients based on these correlations. The effectiveness of the modified assessment model for VTE prediction was also validated. Methods A retrospective study was conducted, involving 51 VTE patients who were hospitalized in the Respiratory Department of Huaian First People’s Hospital from March 2019 to July 2023. These patients were compared with 1,600 non-VTE patients who were discharged during the same period. Clinical data, including medical history and laboratory test results, were retrospectively collected from both groups. The correlation between clinical data and VTE occurrence was analyzed, and highly relevant risk factors were incorporated into the Padua score. The modified Padua risk assessment model was applied to all patients and validated in a validation group. The scores from both the original and modified risk assessment models were compared to evaluate the effectiveness of the modified Padua score. Results Rank sum tests showed significant differences in basic information, such as age, BMI, and length of hospital stay, as well as laboratory tests including mean corpuscular volume, procalcitonin, albumin, alanine aminotransferase, aspartate aminotransferase, urea, and D-dimer (P<0.05). Univariate and multivariate logistic regression analyses revealed that newly identified high-risk factors for VTE included hypoalbuminemia (OR=2.972), blood transfusion (OR=47.035), and mechanical ventilation (OR=6.782) (P<0.05). Receiver operating characteristic curve analysis showed that the sensitivity and specificity of the modified Padua score were higher than those of the original version. The area under the curve (AUC) difference was 0.058, with a Z-test value of 2.442, showing statistical significance (P<0.05). Conclusions The modified Padua score demonstrated superior predictive ability for VTE in hospitalized respiratory patients compared to the original Padua score.
Objective To evaluate proton MR spectroscopy (1H-MRS) for detection of the motor cortex and adjacent brain in amyotrophic lateralsclerosis (ALS) patients with apparent upper motor neuron involvement after olfactory ensheathing cells(OECs) transplantation. Methods From December 2004 to February 2005, 7 patients with clinically definite ALS who could safely undergo MRS were admitted into the perspective study. The neurological status, ALS functional rating scale (ALSFRS), EMG, and 1H-MRS taken before and 2 weeks after operationswere carefully analyzed. The NAA/Cr and Cho/Cr ratios were measured in the cerebral peduncle,genu and posterior limb of the internal capsule, corona radiata and precentral gyrus. Results The ALSFRS in 2 cases mproved obviously whose ALSFRS increased from 30 to 33 and from 29 to 34 respectively. And 5 cases remained stable 2 weeks after OECs transplantation. Statistical analyses for all seven cases showed both theNAA/Cr and Cho/Cr ratios decreased, but in the two cases with ALSFRS improvement the NAA/Cr increased in the certain anatomic position which confirmed the neurological and EMG findings. Conclusion The proton MR spectroscopy is a suitablenoninvasive measure for ALS evaluation. The preliminary study suggests that twoof the seven ALS cases improved apparently shortterm after OECs transplantation. More patients are required for the clinical study and longer followup duration is needed for future research.
Lung nodules are the main manifestation of early lung cancer. So accurate detection of lung nodules is of great significance for early diagnosis and treatment of lung cancer. However, the rapid and accurate detection of pulmonary nodules is a challenging task due to the complex background, large detection range of pulmonary computed tomography (CT) images and the different sizes and shapes of pulmonary nodules. Therefore, this paper proposes a multi-scale feature fusion algorithm for the automatic detection of pulmonary nodules to achieve accurate detection of pulmonary nodules. Firstly, a three-layer modular lung nodule detection model was designed on the deep convolutional network (VGG16) for large-scale image recognition. The first-tier module of the network is used to extract the features of pulmonary nodules in CT images and roughly estimate the location of pulmonary nodules. Then the second-tier module of the network is used to fuse multi-scale image features to further enhance the details of pulmonary nodules. The third-tier module of the network was fused to analyze the features of the first-tier and the second-tier module of the network, and the candidate box of pulmonary nodules in multi-scale was obtained. Finally, the candidate box of pulmonary nodules under multi-scale was analyzed with the method of non-maximum suppression, and the final location of pulmonary nodules was obtained. The algorithm is validated by the data of pulmonary nodules on LIDC-IDRI common data set. The average detection accuracy is 90.9%.