Multimodal data modeling can effectively overcome the problem of insufficient information in a single mode and can greatly improve the performance of model. However, not much progress has been made in quantifying the confidence of neural network models, especially for multimodal fusion models. This paper proposes a method based on embedding, which calculates the local density estimation in the embedding space by calculating the distance between samples, and then calculates the confidence score of the model. The proposed method is scalable and can be used not only for a single modal model, but also for the confidence measurement of multimodal fusion model. In addition, it can also evaluate and quantify the influences of different modal data on the multimodal fusion model.
In order to ensure the stable operation of the miniature gastrointestinal robot inside the human body, a new type of wireless power transmission (WPT) system is presented. To fulfill the requirements of the work energy consumption, posture stability, energy supply stability, and size limitation of the intestinal robot, the design refers to the calculation of electromagnetic field theory and the optimization results of the spindle-shaped three-dimensional receiving coil model which based on the principle of electromagnetic field resonance. Accordingly, the transmitting coil adopts four sets of solenoid coils to generate a uniform alternating magnetic field. With the receiving coil integrated inside, the optimized transmitting coil can generate a stable energy greater than 500 mW for a robot to work through mutual inductance. Relevant experimental verification indicates that performance of the wireless energy supply system satisfies the working requirements of the robot in the complex environment inside the intestine.
The Markov decision process (MDP) model is developed for the patient admission control problem and is extended based on the uniformization method to realize a real-time period-by-period decision process. The classical MDP iterative method is extended, and the two-way iteration algorithm and the two-way threshold iteration algorithm are proposed to solve the new model. Numerical experiments are conducted based on the data from the resuscitation room of a large hospital in Shanghai. The results show that the proposed method can improve the admission rate of critical patients and enhance the medical service level of the hospital compared with the existing decision method.
The accurate classification of shape, edge, and internal features of pulmonary nodules can not only assist the radiologists in their daily diagnosis, but also improve the writing efficiency of imaging reports. This paper proposes a multi-task classification model based on long-short term memory (LSTM) and attention structure, which merges the shared features among different classification tasks through attention mechanism to improve the feature extraction performance of the current task. The classifier based on temporal sequence LSTM structure can effectively screen the shared features and improve the efficiency of information transmission in the multi-task model. Experimental results show that compared with the traditional multi-task structure, the proposed model can achieve better classification results on the public dataset LIDC-IDRI, and assist doctors to quickly obtain nodule characteristics.
In order to study the spread characteristics of virus droplets in the bus and predict the risk probability of airborne virus infecting passengers in the bus, a numerical model of the bus with the refrigeration and air conditioning on in summer is established based on the numerical simulation of computational fluid dynamics (CFD) in combination with the Wells-Riley equation. Through the analysis of the characteristics of the flow field in the bus in combination with the Lagrangian method, the spread process of the droplets produced by the cough of the virus carrier in the bus is calculated. The infection risk of virus-carrying droplets to the passengers in the bus in the four ventilation mades is analyzed. It is found that the longitudinal airflow in the bus is a key factor affecting the spread of droplets. Compared with the asymmetrically arranged circular air outlets, the longitudinal airflow in the bus can be reduced by the use of slit-type air outlets. Only 6% of passengers have a higher than 5% probability of contracting diseases in the bus with displacement ventilation, which is more effective in reducing the risk of droplet-borne infection. The research results can provide guidance for the structural design of bus air supply systems and reduce the risk of droplet-borne infections.
As an implantable medical device, the transcutaneous wireless energy transmission (TET) system of biaxial actuated artificial anal sphincter (BAAS) after implantation in vivo is affected by the hyperplasia of the parcel. The coupling performance is obviously lower than that of in vitro experiments, which reduces the practical value in medical treatment. Based on the experimental results of the TET system of BAAS, mainly aimed at the problem of transmission angle caused by the hyperplasia of the parcel, the existing TET system is analyzed and studied, and the relation curve between transmission angle and transmission efficiency are obtained. The experimental results show that when the transmission distance is 30 mm, the transmission efficiency reaches more than 60%. When the transmission distance is 30 mm and the transmission angle is 20°, the average transmission efficiency of the experiment can reach 66.65%. When the transmission angle is 30°, the average transmission efficiency of the experiment can reach 59.96%. The maximum of charging temperature of the transmitting coil is 23 ℃, which will not cause low temperature scald to the human body. It can meet the performance requirements of the BAAS system and achieve the purpose of long distance and high efficiency transmission.
Mandibular angle osteotomy is a popular craniofacial plastic surgery in recent years. Usually, preoperative planning of mandibular angle osteotomy is completed by an experienced doctor, which is cumbersome and time-consuming. In order to improve the efficiency of osteotomy planning, a design method of mandibular angle osteotomy plane based on point cloud semantic segmentation network is proposed. After three-dimensional reconstruction of the skull computer tomography (CT) scan data, the three-dimensional model of the mandible is converted into point cloud data through uniform sampling. The resection area of the mandible is predicted by the proposed algorithm, which is used to calculate the mandibular angle osteotomy plane. The proposed semantic segmentation network mainly includes 2 parts: a local feature extraction layer based on the attention mechanism, which is used to extract fine-grained local structure information, and a non-local feature extraction layer based on Transformer, which is used to extract the global context information of the point cloud. On the constructed mandible semantic segmentation data set, the proposed algorithm is compared with other point cloud semantic segmentation algorithms. The results show that the proposed algorithm can achieve the best prediction of the mandibular angle resection area, which is better than current common point cloud semantic segmentation algorithms.
The electromagnetic interference (EMI) from pulsed electromagnetic field (PEMF) on pacemakers is unignorable in modern power grids and healthcare environments, but there is limited study on the interaction mechanisms and protective measures. In this paper, an in-vitro human chest model for pacemaker implantation is made by using pork tissues immersed in 0.9% sodium chloride solution. The effect of PEMFs generated by the switching actions of common electrical equipment and low-frequency medical equipment on pacemakers is simulated by using fast-front current sources. The pulse forming line theory is employed for analyzing the waveform compression of PEMFs in human thoracic cavity. Further, the parameterized bio-electromagnetic transient model of pacemaker in combination with biological tissues is established in finite element software. The results show that pacemaker malfunctions including pacing inhibition and P pulmonale occur under PEMF. The “Window effect” in subcutaneous pouch under PEMF is found by changing the winding of pacemaker leads in the pouch. Based on the research finding, a protective strategy by using composite materials to shield the window area is proposed. The theoretical feasibility of this protective measure is confirmed by simulation, where the intensity of pacemaker EMI could be reduced by 80 dB when the composite materials shielding is used. Finally, a safe distance is developed for pacemaker wearers in electrical and medical environments.
