Improved HHT-microstate analysis of EEG in nicotine addicts.

Background: Substance addiction is a chronic disease which causes great harm to modern society and individuals. At present, many studies have applied EEG analysis methods to the substance addiction detection and treatment. As a tool to describe the spatio-temporal dynamic characteristics of large-scale electrophysiological data, EEG microstate analysis has been widely used, which is an effective method to study the relationship between EEG electrodynamics and cognition or disease. Methods: To study the difference of EEG microstate parameters of nicotine addicts at each frequency band, we combine an improved Hilbert Huang Transformation (HHT) decomposition with microstate analysis, which is applied to the EEG of nicotine addicts. Results: After using improved HHT-Microstate method, we notice that there is significant difference in EEG microstates of nicotine addicts between viewing smoke pictures group (smoke) and viewing neutral pictures group (neutral). Firstly, there is a significant difference in EEG microstates at full-frequency band between smoke and neutral group. Compared with the FIR-Microstate method, the similarity index of microstate topographic maps at alpha and beta bands had significant differences between smoke and neutral group. Secondly, we find significant class × group interactions for microstate parameters at delta, alpha and beta bands. Finally, the microstate parameters at delta, alpha and beta bands obtained by the improved HHT-microstate analysis method are selected as features for classification and detection under the Gaussian kernel support vector machine. The highest accuracy is 92% sensitivity is 94% and specificity is 91%, which can more effectively detect and identify addiction diseases than FIR-Microstate and FIR-Riemann methods. Conclusion: Thus, the improved HHT-Microstate analysis method can effectively identify substance addiction diseases and provide new ideas and insights for the brain research of nicotine addiction.

A deep learning model for the differential diagnosis of benign and malignant salivary gland tumors based on ultrasound imaging and clinical data.

Background: The preoperative differentiation between benign parotid gland tumors (BPGTs) and malignant parotid gland tumors (MPGTs) is of great significance for therapeutic decision-making. Deep learning (DL), an artificial intelligence algorithm based on neural networks, can help overcome inconsistencies in conventional ultrasonic (CUS) examination outcomes. Therefore, as an auxiliary diagnostic tool, DL can support accurate diagnosis using massive ultrasonic (US) images. This current study developed and validated a DL-based US diagnosis for the preoperative differentiation of BPGT from MPGT. Methods: A total of 266 patients, including 178 patients with BPGT and 88 patients with MPGT, were consecutively identified from a pathology database and enrolled in this study. Ultimately, considering the limitations of the DL model, 173 patients were selected from the 266 patients and divided into 2 groups: a training set, and a testing set. US images of the 173 patients were used to construct the training set (including 66 benign and 66 malignant PGTs) and testing set (consisting of 21 benign and 20 malignant PGTs). These were then preprocessed by normalizing the grayscale of each image and reducing noise. Processed images were imported into the DL model, which was then trained to predict the images from the testing set and evaluated for performance. Based on the training and validation datasets, the diagnostic performance of the 3 models was assessed and verified using receiver operating characteristic (ROC) curves. Ultimately, before and after combining the clinical data, we compared the area under the curve (AUC) and diagnostic accuracy of the DL model with the opinions of trained radiologists to evaluate the application value of the DL model in US diagnosis. Results: The DL model showed a significantly higher AUC value compared to doctor 1 + clinical data, doctor 2 + clinical data, and doctor 3 + clinical data (AUC =0.9583 vs. 0.6250, 0.7250, and 0.8025 respectively; all P<0.05). In addition, the sensitivity of the DL model was higher than the sensitivities of the doctors combined with clinical data (97.2% vs. 65%, 80%, and 90% for doctor 1 + clinical data, doctor 2 + clinical data, and doctor 3 + clinical data, respectively; all P<0.05). Conclusions: The DL-based US imaging diagnostic model has excellent performance in differentiating BPGT from MPGT, supporting its value as a diagnostic tool for the clinical decision-making process.

Segmentation of retinal vessels based on MRANet.

The segmentation of retinal vessel takes a crucial part in computer-aided diagnosis of diseases and eye disorders. However, the insufficient segmentation of the capillary vessels and weak anti-noise interference ability make such task more difficult. To solve this problem, we proposed a multi-scale residual attention network (MRANet) which is based on U-Net network. Firstly, to collect useful information about the blood vessels more effectively, we proposed a multi-level feature fusion block (MLF block). Then, different weights of each fused feature are learned by using attention blocks, which can retain more useful feature information while reducing the interference of redundant features. Thirdly, multi-scale residual connection block (MSR block) is constructed, which can better extract the image features. Finally, we use the DropBlock layer in the network to reduce the network parameters and alleviate network overfitting. Experiments show that based on DRIVE, the accuracy rate and the AUC performance value of our network are 0.9698 and 0.9899 respectively, and based on CHASE_DB1 dataset, they are 0.9755 and 0.9893 respectively. Our network has a better segmentation effect compared with other methods, which can ensure the continuity and completeness of blood vessel segmentation.

[Application of a parallel branches network based on Transformer for skin melanoma segmentation].

Cutaneous malignant melanoma is a common malignant tumor. Accurate segmentation of the lesion area is extremely important for early diagnosis of the disease. In order to achieve more effective and accurate segmentation of skin lesions, a parallel network architecture based on Transformer is proposed in this paper. This network is composed of two parallel branches: the former is the newly constructed multiple residual frequency channel attention network (MFC), and the latter is the visual transformer network (ViT). First, in the MFC network branch, the multiple residual module and the frequency channel attention module (FCA) module are fused to improve the robustness of the network and enhance the capability of extracting image detailed features. Second, in the ViT network branch, multiple head self-attention (MSA) in Transformer is used to preserve the global features of the image. Finally, the feature information extracted from the two branches are combined in parallel to realize image segmentation more effectively. To verify the proposed algorithm, we conducted experiments on the dermoscopy image dataset published by the International Skin Imaging Collaboration (ISIC) in 2018. The results show that the intersection-over-union (IoU) and Dice coefficients of the proposed algorithm achieve 90.15% and 94.82%, respectively, which are better than the latest skin melanoma segmentation networks. Therefore, the proposed network can better segment the lesion area and provide dermatologists with more accurate lesion data.

Diagnosis of rectal cancer based on the Xception-MS network.

Objective. Accurate T staging of rectal cancer based on ultrasound images is convenient for doctors to determine the appropriate treatment. To effectively solve the problems of low efficiency and accuracy of traditional methods for T staging diagnosis of rectal cancer, a deep-learning-based Xception-MS diagnostic model is proposed in this paper.Approach. The proposed diagnostic model consists of three steps. First, the model preprocesses rectal cancer images to solve the problem of data imbalance and deficiency of sample size, and reduces the risk of model overfitting. Second, a new Xception-MS network with stronger feature extraction capability, which is a combination of the Xception network and MS module, is proposed. The MS module is a new function module that can more effectively extract multi-scale information from rectal cancer images. In addition, to solve the deficiency of the small sample size of rectal cancer, the proposed network is combined with transfer learning technology. At last, the output layer of the network is modified, in which the global average pooling and a fully connected softmax layer are employed to replace the original ones, and then the rectal cancer 4 classification (T1, T2, T3, T4 staging) is output.Main results. Experiments of rectal cancer T staging are conducted on a dataset of 1078 rectal cancer images in 4 categories collected from the Department of Colorectal Surgery of the Third Affiliated Hospital of Kunming Medical University. The experimental results show that the accuracy, precision, recall andF1 values obtained by the model are 94.66%, 94.70%, 94.65%, and 94.67%, respectively.Significance. The experimental results show that our model is superior to the existing classification models, can effectively and automatically classify ultrasound images of rectal cancer, and can better assist doctors in the diagnosis of rectal cancer.

A Multimodal Classification Architecture for the Severity Diagnosis of Glaucoma Based on Deep Learning.

Glaucoma is an optic neuropathy that leads to characteristic visual field defects. However, there is no cure for glaucoma, so the diagnosis of its severity is essential for its prevention. In this paper, we propose a multimodal classification architecture based on deep learning for the severity diagnosis of glaucoma. In this architecture, a gray scale image of the visual field is first reconstructed with a higher resolution in the preprocessing stage, and more subtle feature information is provided for glaucoma diagnosis. We then use multimodal fusion technology to integrate fundus images and gray scale images of the visual field as the input of this architecture. Finally, the inherent limitation of convolutional neural networks (CNNs) is addressed by replacing the original classifier with the proposed classifier. Our architecture is trained and tested on the datasets provided by the First Affiliated Hospital of Kunming Medical University, and the results show that the proposed architecture achieves superior performance for glaucoma diagnosis.

EEG microstate in obstructive sleep apnea patients.

Obstructive sleep apnea (OSA) is a common sleep respiratory disease. Previous studies have found that the wakefulness electroencephalogram (EEG) of OSA patients has changed, such as increased EEG power. However, whether the microstates reflecting the transient state of the brain is abnormal is unclear during obstructive hypopnea (OH). We investigated the microstates of sleep EEG in 100 OSA patients. Then correlation analysis was carried out between microstate parameters and EEG markers of sleep disturbance, such as power spectrum, sample entropy and detrended fluctuation analysis (DFA). OSA_OH patients showed that the microstate C increased presence and the microstate D decreased presence compared to OSA_withoutOH patients and controls. The fifth microstate E appeared during N1-OH, but the probability of other microstates transferring to microstate E was small. According to the correlation analysis, OSA_OH patients in N1-OH showed that the microstate D was positively correlated with delta power, and negatively correlated with beta and alpha power; the transition probability of the microstate B → C and E → C was positively correlated with alpha power. In other sleep stages, the microstate parameters were not correlated with power, sample entropy and FDA. We might interpret that the abnormal transition of brain active areas of OSA patients in N1-OH stage leads to abnormal microstates, which might be related to the change of alpha activity in the cortex.

A segmentation method combining probability map and boundary based on multiple fully convolutional networks and repetitive training.

Cell nuclei image segmentation technology can help researchers observe each cell's stress response to drug treatment. However, it is still a challenge to accurately segment the adherent cell nuclei. At present, image segmentation based on a fully convolutional network (FCN) is attracting researchers' attention. We propose a multiple FCN architecture and repetitive training (M-FCN-RT) method to learn features of cell nucleus images. In M-FCN-RT, the multiple FCN (M-FCN) architecture is composed of several single FCNs (S-FCNs) with the same structure, and each FCN is used to learn the specific features of image datasets. In this paper, the M-FCN contains three FCNs; FCN1-2, FCN3 and FCNB. FCN1-2 and FCN3 are respectively used to learn the spatial features of cell nuclei for generating probability maps to indicate nucleus regions of an image; FCNB (boundary FCN) is used to learn the edge features of cell nuclei for generating the nucleus boundary. For the training of each FCN, we propose a repetitive training (RT) method to improve the classification accuracy of the model. To segment cell nuclei, we finally propose an algorithm combining the probability map and boundary (PMB) to segment the adherent nuclei. This paper uses a public opening nucleus image dataset to train, verify and evaluate the proposed M-FCN-RT and PMB methods. Our M-FCN-RT method achieves a high Dice similarity coefficient (DSC) of 92.11%, 95.64% and 87.99% on the three types of sub-datasets respectively for probability maps. In addition, segmentation experimental results show the PMB method is more effective and efficient compared with other methods.

Ultrasound speckle reduction based on fractional order differentiation.

PURPOSE: Ultrasound images show a granular pattern of noise known as speckle that diminishes their quality and results in difficulties in diagnosis. To preserve edges and features, this paper proposes a fractional differentiation-based image operator to reduce speckle in ultrasound. METHODS: An image de-noising model based on fractional partial differential equations with balance relation between k (gradient modulus threshold that controls the conduction) and v (the order of fractional differentiation) was constructed by the effective combination of fractional calculus theory and a partial differential equation, and the numerical algorithm of it was achieved using a fractional differential mask operator. RESULTS: The proposed algorithm has better speckle reduction and structure preservation than the three existing methods [P-M model, the speckle reducing anisotropic diffusion (SRAD) technique, and the detail preserving anisotropic diffusion (DPAD) technique]. And it is significantly faster than bilateral filtering (BF) in producing virtually the same experimental results. CONCLUSIONS: Ultrasound phantom testing and in vivo imaging show that the proposed method can improve the quality of an ultrasound image in terms of tissue SNR, CNR, and FOM values.

[Development of a lung cancer image database and visualization toolkit].

Lung cancer is the most common tumor and one of the malignant tumors with the lowest livability after diagnosis, as is known so far. Large-scale image database is the foundation of developing computer-aided diagnosis methods, education and training in lung cancer diagnosis to improve medical diagnostic efficiency and to reduce the doctors' burden. In this study, aiming at improving the low data storage efficiency and solving the lacking of tool for data visualization and data retrieval existing in the use of traditional Lung Image Database Consortium (LIDC) from the lung cancer database, we developed a new lung cancer image database platform including an improved data model, a data integration tool, an image and annotation visualization tool and a data retrieving component. Firstly, the data format in LIDC was analyzed and an improved information model was provided to manage and manipulate large amount data stored in it. Next, some tools such as data integration component, DICOM, image and annotation visualization tool, and data query were designed and implemented. The study demonstrated that the lung cancer image database platform had the capacity of data collection, visualization, and query, and could promote diagnose lung cancer research.