JICE

Showing 25–36 of 85 results

  • Association Analysis of primary Liver Cancer based on Apriori Algorithm

    Open Access

    Abstract: This paper briefly describes the Apriori algorithm for primary liver cancer data, and then performs data preprocessing based on the characteristics of primary liver cancer patients, including data import and extraction, and embedding the algorithm into primary liver cancer. The implementation of clinical warning. After that, the Apriori algorithm is used to realize the association of the data association rules of primary liver cancer, and the internal valuable association rules are obtained, which provides suggestions for improving the doctor’s remediation and prevention, so as to prevent the occurrence and reduction of primary liver cancer. The incidence of primary liver cancer has important practical significance.

    Keywords: Apriori algorithm; primary liver cancer; Association rules; association analysis

  • Development of Drone Detecting Free Parking Space for Car Parking Guidance

    Open Access

    Abstract: In the course of transportation of medical drugs, the requirements for temperature and humidity are high. We should strictly control it within a certain range, otherwise the deterioration of drugs will occur, which will lead to serious consequences. Therefore, a cold chain transport vehicle with temperature and humidity sensors should be used in medical transportation. This paper designs a wireless monitoring system based on LoRa technology, which can collect the data of temperature and humidity sensors on the transport vehicle in real time, and transmit them to the background through the LoRa gateway. Solve the technical problems of high power consumption and data upload failure. A wireless cold chain monitoring system with high data accuracy, strong expansibility and convenient networking is realized ensure the quality of drugs quantity safety.

    Keywords: Drone; ardrone 2.0; template matching; parking lot

  • Research on Robot Path Planning Based on Dijkstra and Ant Colony Optimization

    Open Access

    Abstract: This paper studied on the path planning problem in known environments. According to Dijkstra algorithm and ant colony optimization (ACO), a hybrid algorithm to search the path was designed. Based on the environment model, constructed by using visual graph method, Dijkstra algorithm was used for initial path planning. Then the ACO was improved and used to optimize the initial path to minimize the path of the robot. The simulation on MATLAB showed that the path planning algorithm based on Dijkstra-ACO has higher efficiency of path search and good effect of path planning, and the algorithm is effective and feasible.

    Keywords: Dijkstra Algorithm, Ant Colony Optimization (ACO), Path Planning

  • A Medicine Cold Chain Monitor System Based on LoRa Wireless Technology

    Open Access

    Abstract: In the course of transportation of medical drugs, the requirements for temperature and humidity are high. We should strictly control it within a certain range, otherwise the deterioration of drugs will occur, which will lead to serious consequences. Therefore, a cold chain transport vehicle with temperature and humidity sensors should be used in medical transportation. This paper designs a wireless monitoring system based on LoRa technology, which can collect the data of temperature and humidity sensors on the transport vehicle in real time, and transmit them to the background through the LoRa gateway. Solve the technical problems of high power consumption and data upload failure. A wireless cold chain monitoring system with high data accuracy, strong expansibility and convenient networking is realized ensure the quality of drugs quantity safety.

    Keywords: LoRa, Things of Internet, Cold Chain, Sensor, Temperature and Humidity

  • Research on Text Classification Method based on PTF-IDF and Cosine Similarity

    Open Access

    Abstract: Text classification is a foundational task in many NLP applications. The text classification task in the era of big data faces new challenges. We propose a Promoted TF-IDF (Promoted-TF-IDF) and cosine similarity method for text classification. In our model, with the pre-trained word segmentation tool, we apply PTF-IDF method to judge which words play key roles in text classification to capture the key components in category. We also apply Cosine Similarity algorithm to judge similarity between text and category. We conduct experiments on commonly used datasets. The experimental results show that the proposed method outperforms the state-of-the-art methods on several datasets.

    Keywords: Text classification, TF-IDF, Computer Application, Natural Language Processing

  • An attention mechanism-based multi-scale network crowd density estimation algorithm

    Open Access

    Abstract—It is becoming more and more important to calculate the people number in terms of the requirement for the safety management, because that the crowd gathering scenes are common whether or not it is daily urban traffic or some special gatherings. Calculating the people number in high-density crowd is a very difficult challenge due to the diversity of ways people appear in crowded scenes. This paper proposes a multi-branch network which combines the dilated convolution and attention mechanism. By combining dilated convolution, the context information of different scales of the crowd image are extracted. The attention mechanism is introduced to make the network pay more attention to the position of the head of the crowd and suppress the background noise, so as to obtain a higher quality density map. Then add all the pixels in the density map to get the total people number. Through a large number of experiments, this network can better provide effective crowd density estimation features and improve the dissimilarity of density map distribution, which has stronger robustness.

    Keywords: Attention mechanism; Crowd density estimation; Dilated convolutional

  • Vehicle Detection Method Based on ADE-YOLOV3 Algorithm

    Open Access

    Abstract: Aiming at the problem of repeated detection of YOLOV3 algorithm in vehicle detection, the ADE-YOLOV3 vehicle detection algorithm is proposed. The algorithm uses K-means clustering algorithm to determine the number of target candidate frames and aspect ratio according to the inherent width and height characteristics of the vehicle. Then, according to the results obtained by clustering, the anchor parameters are reset, which makes the ADE-YOLOV3 network have certain pertinence in vehicle detection. Finally, the migration learning method is used to improve the network structure, and the optimal weight model is obtained, which improves the training precision of the model. The experimental results show that compared with the original YOLOV3 method, the mAP is increased from 91.4% to 95%, the repeated detection rate is reduced from 5.6% to 2.1%,and the detection speed by 50fps. The detection accuracy is improved and the problem of repeated detection is effectively avoided.

    Keywords: Vehicle Detection; Deep Learning; YOLOv3; K-means; Migration Learning

  • Construction of Student model based on BP neural network

    Open Access

    Abstract: With the development of personalized learning, the construction of student models is becoming more and more important. At present, there are still problems in the student model that the characteristics are single and the indicators of each dimension are not clear. In this paper, learners will be analyzed from the perspective of student characteristics. And BP (Back Propagation) neural network algorithm will be used to establish a personalized student model. This paper first constructs the feature system of the student model from six dimensions. Secondly, the initial data is obtained through questionnaire survey, and the data is initialized to obtain 30 feature vectors as input to BP neural network. The output of the network is a learner type, which is divided into 36 categories. The construction of the student model will have certain practical significance for realizing the effectiveness of personalized education in distance education.

    Keywords: Student Model, Personalized Learning, BP Neural Network, Student Characteristics

  • A Game-based Upper Limb AROM Measurement System for Older Adults

    Open Access

    Abstract:  At present, the aging problem is becoming more and more serious, which undoubtedly causes people’s attention to the life quality of the elderly. Active range of motion (AROM) is an important index to judge the ability of daily functional activities of the elderly. Therefore, it is very important to obtain the AROM data of the elderly quickly and accurately. The traditional method of measuring the range of motion (ROM) is to measure the passive range of motion (PROM) with a goniometer by nurses, which cannot accurately reflect the elderly’s active ability in the daily life. In order to overcome the shortcomings of traditional methods, the AROM measurement system has been developed. However, in the current AROM measurement system, the elderly is guided to several specific postures, ignoring the positive role of interest in stimulating the elderly to participate in ROM measurement. In this paper, we propose a game-based upper limb AROM measurement system. In the system, the joint coordinates of the player are measured by the depth image sensor, and the AROM is calculated automatically and objectively by using the coordinates. From the experimental results, the average flection value measured by our system is 21.6 degrees larger than that measured by the goniometer. The average abduction value measured by our system is 24.1 degrees larger than that measured by goniometer. This result means that the elderly can stretch better through a game in our measurement system. In order to encourage the elderly to participate more in the game, a questionnaire survey was conducted on the elderly’s views of the game. From the analysis results, we find that the merit factor and design factor of the game have a great impact on the players’ experience in the game. The research results provide us with ideas for the improvement of the measurement system in the

  • Model for Non-contact Blood Pressure Measurement Using the Facial Feature Amount based on Amplitude and Phase Analysis

    Open Access

    Abstract:  To monitor the daily blood pressure, developing a non-contact method for measuring blood pressure is necessary. In a previous study, we proposed a novel method that described the vascular structure of an entire human face as an electric circuit based on amplitude and phase analyses using visible and thermal images of the face. However, the model developed by that method did not consider the order of blood flow because the model applied at random the extracted features. In the present study, it considers a model that incorporates the order of blood flow utilizing amplitude and phase analyses. As a result, the estimated accuracy is improved by considering the order of blood flow. Higher accuracy requires more detailed vascular structure information of the face. It concluded that the facial feature, which is related to the blood pressure, can be obtained by the CEOF analysis.

  • Construction of an individual model for estimating blood pressure using independent components of facial skin temperature considering time variation

    Open Access

    Abstract:  The objective of this study was to construct an individual model for estimating blood pressure (BP) using independent components of facial skin temperature considering time variation. In our previous study, an individual model was constructed for estimating BP by applying independent component analysis to facial skin temperature of each subject. However, in this previous study, time variations in facial skin temperature were not considered. Facial skin temperature is assumed to be related to blood flow over time as blood helps in transporting heat in the body. Therefore, the accuracy of the BP estimation model can be expected to improve if the variation of facial skin temperature caused by blood flow over time considered in addition to only the facial skin temperature. Consequently, the accuracy of the proposed model with the aforementioned considerations was found to be better than that of without these considerations. Therefore, the BP was accurately estimated using the proposed approach.  

  • Research on Gesture Based on GA – SVM

    Open Access

    Abstract:  Surface electromyography (sEMG) is a kind of weak electrical signal generated by muscle activity, which contains information of gesture and is widely used in prosthetic control, rehabilitation and medical treatment. The difference between different motion patterns can be reflected by the different sEMG characteristics, so the recognition of human motion can be studied. Four time-domain features including absolute mean value, waveform length, zero-crossing number and root mean square value were extracted from the double Myo arm-ring data set in Ninapro benchmark database. Classification and identification were performed by using the Support Vector Machine (SVM) optimized by Genetic Algorithms (GA). Experimental results showed that the optimized SVM classification had a better effect.