Accurate cortical surface parcellation is essential for elucidating brain organizational principles, functional mechanisms, and the neural substrates underlying higher cognitive and emotional processes. However, the cortical surface is a highly folded complex geometry, and large regional variations make the analysis of surface data challenging. Current methods rely on geometric simplification, such as spherical expansion, which takes hours for spherical mapping and registration, a popular but costly process that does not take full advantage of inherent structural information. In this study, we propose an Attention-guided Deep Graph Convolutional network (ADGCN) for end-to-end parcellation on primitive cortical surface manifolds. ADGCN consists of a deep graph convolutional layer with a symmetrical U-shaped structure, which enables it to effectively transmit detailed information of the original brain map and learn the complex graph structure, help the network enhance feature extraction capability. What’s more, we introduce the Squeeze and Excitation (SE) module, which enables the network to better capture key features, suppress unimportant features, and significantly improve parcellation performance with a small amount of computation. We evaluated the model on a public dataset of 100 artificially labeled brain surfaces. Compared with other methods, the proposed network achieves Dice coefficient of 88.53% and an accuracy of 90.27%. The network can segment the cortex directly in the original domain, and has the advantages of high efficiency, simple operation and strong interpretability. This approach facilitates the investigation of cortical changes during development, aging, and disease progression, with the potential to enhance the accuracy of neurological disease diagnosis and the objectivity of treatment efficacy evaluation.
There is a lack of study on fault detection methods of medical equipment at home and abroad. The main reason is that the research of fault features is diverse and not systematic. This paper aims to propose a fault recognition method for medical equipment combining the electrical performance parameter features with fault events. First, it treats the equipment as a whole system, setting up the analysis model. Then, we are going to analyze the signal for indicator. This paper chooses the multi-index electrical performance parameters (MEPP) method for the fault identification an indicator. It is proved that the electrical performance signal can evaluate the status of equipment. Thus, it can also be used to recognize the fault or other working statuses. Then, the features of current, voltage, and power are studied exhaustively using a mathematical model. After that, the weight of each parameter feature in any specific event will be determined according to the influence of each parameter feature on fault events. At that time, the recognition method basically realizes the correlation between multi-index features and fault events through weight. Next, the above method needs to be verified in the experiment. This paper chooses six monitors for setting the rules of normal status. The normal status is the baseline for fault identification. Then, feature intervals of other faults are established around this reference. Finally, each feature interval will be constantly adjusted to meet the preset recognition rate and updated to the rules in the subsequent measurement. In this paper, 10 monitors are selected as samples to update a set of basic fault judgment rules based on MEPP, and by adjusting the overlapping interval, the fault recognition rate reaches more than 90% in this study. To sum up, this paper uses the MEPP method to find out the relationship of features of current, voltage, and power with fault events. It will become a new direction for fault recognition studies on electrical medical equipment and other device.
16 April 2025 in IEEE Transactions on Instrumentation and Measurement
Currently, electrical capacitance tomography (ECT) is limited to time-differential imaging for monitoring dynamic alterations in cerebral hemorrhage. However, the inherent constraint of this approach renders it unsuitable for rapid hemorrhage detection, as it requires a reference measurement from a non-hemorrhaging brain. To address this limitation, this study proposes a novel approach of frequency-differential ECT (FDECT) for cerebral hemorrhage imaging in practice. The method entails identification of a frequency range wherein the permittivity variation of cerebral blood with frequency is much greater than the variation of other brain tissues. Within this identified range, two optimal frequencies are selected, and the permittivity difference at these two frequencies are used for imaging. With this method, cerebral hemorrhage is highlighted and other brain tissue is suppressed, thereby achieving the absolute distribution of cerebral hemorrhage and eliminating the need for non-hemorrhagic baseline data. Simulation results demonstrate that FDECT imaging quality correlates directly with the frequency-dependent permittivity difference between the target and background media, thereby validating FDECT's theoretical basis and highlighting the critical role of optimal frequency selection. Before conducting in vitro animal imaging, we analyzed the dielectric spectra of ex vivo sheep blood, pig fat, and pig brain tissue to identify the optimal frequency range for differentiating blood from these tissues. In vitro experiments confirmed that FDECT with the optimal frequencies effectively images blood within pig fat or brain tissue, contrasting with the suboptimal results from non-ideal frequencies. Although essential for FDECT success, optimal frequency pairing does not eliminate the higher noise levels in FDECT images, largely due to the background brain tissue's frequency-dependent dielectric characteristics. To mitigate this inherent limitation and improve imaging quality, we intend to implement a three-frequency FDECT approach, reducing background tissue interference.
27 November 2023 in Journal of Artificial Intelligence and Technology
Achieving accurate segmentation of brain tumors in Magnetic Resonance Imaging (MRI) is important for clinical diagnosis and accurate treatment, and the efficient extraction and analysis of MRI multimodal feature information is the key to achieving accurate segmentation. In this paper, we propose a multimodal information fusion method for brain tumor segmentation, aimed at achieving full utilization of multimodal information for accurate segmentation in MRI. In our method, the semantic information processing module (SIPM) and Multimodal Feature Reasoning Module (MFRM) are included: (1) SIPM is introduced to achieve free multiscale feature enhancement and extraction; (2) MFRM is constructed to process both the backbone network feature information layer and semantic feature information layer. Using extensive experiments, the proposed method is validated. The experimental results based on BraTS2018 and BraTS2019 datasets show that the method has unique advantages over existing brain tumor segmentation methods.