Wind power generation is an important power generation method as one of renewable energy. It is necessary to reduce accidents that occur at wind power plants and to use them continuously. In Japan, there have been many reports of damage to wind turbine blades due to lightning strikes due to Japan's unique climate. In order to reduce accidents, lightning induction using CFRP is proposed. In this study, we aimed at the adaptation of the lightning strike induction mechanism to the actual machine and examined its influence by the simulation experiment considering the environmental factors.

抄録全体を表示

This study proposes a novel data augmentation method to improve the accuracy of appearance inspection by machine learning, by combining GANs and the cut-and-paste method. When employing this method and training the model with the augmented dataset, the defect detection ratio improved by 2.0pt compared to the case without augmentation. Moreover, the system developed in this study enables automatic annotation of each training data, unlike the conventional supervised learning method that requires manual annotation for each data. This automated system effectively reduces training setup time, human workload, and human variability.

抄録全体を表示

The risk of developing esophageal cancer from Barrett’s Esophagus (BE) is associated with its size. However, the study to measure BE area is not advanced well. In this paper, we propose the method to measure BE area from one endoscopic image with camera distance. First, we assume that the esophagus is a cylinder and the camera is a pinhole camera because 3D reconstruction of no-feature points object like an esophagus is difficult. And then, we get BE area indirectly by calculating the surface area of this virtual cylinder. In measurement experiments, we used USB camera and cylinder phantoms to evaluate the accuracy of the proposed method. The errors were about 4-6% in the bottom area, about 10% in the side area, and about 1±2.5 mm in the phantom’s height measurement.

抄録全体を表示

In this paper, we constructed a system for confirming the open/close status of sluice gates by using inexpensive and low-power sensors that guarantee data accuracy within the allowable range. As a verification test, we constructed a system for monitoring the opening and closing of the sluice gate at the Uchikawa sluice gate in Tsuruoka City, Yamagata Prefecture. As a result, we completed the construction of a sluice gate opening/closing status confirmation system that can be operated stably at a low cost and with low power consumption. This information can be confirmed on the web and can be viewed by Yamagata Prefecture disaster prevention officials.

抄録全体を表示

The pearl pigment emits pearly colors, and its hue differs depending upon the view direction. The authors estimate that this optical feature could apply to the artifact metrics. (1) Two types of optical characteristic information exist in an artifact (a solid epoxy resin) painted with the pigment and were reflected in two types of images (visible images and near-infrared images). (2) Each image from the same image group has a uniqueness when we calculated a correlation ratio using a phase-only correlation method. Hence, we could say that the pearl pigment is suitable for applications to the artifact metrics intended for the solid epoxy resin.

抄録全体を表示

In previous research, a CAPTCHA consisted of one sentence with typoglycemia and multiple pictures was proposed. However, questions presented to research subjects were manually made. Hence, the questions should be created automatically for practical use. Authors have proposed an automatic question making process and developed a software of creating the questions automatically. In our experiment with 50 research subjects, correct answer rate and answer time were the same as the results that the previous study has shown. It means that out process and software have almost reached levels of the results of the previous study.

抄録全体を表示

We have proposed environment adaptation of software that automatically converts normal code suitable for CPU into code suitable for GPU, FPGA and so on depending on the deployment environment and enables high-performance operation. In this paper, we study the overall processing performed for reconfiguration during operation, and newly propose resource amount reconfiguration, measure the processing time for reconfiguration during operation, and discuss reconfiguration timings.

抄録全体を表示

Various viruses have been raging in recent years. Even coronaviruses, which have been decreasing in the number of infections, have not been eliminated. Currently, when disinfection of everyday objects is required, we would like to propose this automatic cleaning system for smartphone screens. In this system, the screen is cleaned by simply placing the smartphone in the grabbing section. The conveyor belt in the grabbing section rotates, and after the smartphone is grabbed, the screen is cleaned by clamping it between the towels in the upper section. In addition, the towel is sprayed with alcohol to disinfect not only the screen of dirt, but also invisible viruses. This system eliminates the time and effort required to clean smartphone screens, which can be a hassle in daily life.

抄録全体を表示

In recent years, the increasing prominence of cyber-attacks has highlighted the critical need for enhanced security in industrial control systems and mobility network systems. These systems, responsible for controlling production lines and plants, are becoming primary targets for cyber threats. In this paper, we propose a method to maintain the stable operation of control systems when cyber-attack anomalies are detected, while simultaneously preventing the spread of damage based on the detected anomalous activities. By integrating real-time anomaly detection with dynamic network zoning, our method aims to promptly detect and isolate any malicious activities or unauthorized devices in the network. In this paper, we demonstrate the effectiveness of our anomaly detection method using security units through experiments in an evaluation environment simulating an industrial control system field network. Furthermore, we show that the implementation of dynamic zoning can effectively mitigate the impact of potential attacks. Our research contributes a robust security solution designed with the unique characteristics and requirements of control systems in mind, offering a promising direction for future advancements in the field.

抄録全体を表示

This paper proposes CNN training using only invariant information clustering (IIC) and non-defective images for visual inspection of industrial products. In the field of visual inspection, there is a problem that there are fewer defective images than non- defective images. To solve this problem, CNN is trained as an encoder using invariant information clustering, and only non-defective images. One-Class Support Vector Machine (OCSVM) was utilized to classify the feature vectors obtained from the encoder. Defective images and non-defective images are classified by detecting values that deviate from the features of the non-defective images. A typical supervised learning model, a model using a variational autoencoder (VAE), and an IIC model are compared. The results confirm that the IIC model that does not use defective images is equivalent to a supervised learning model that uses a small number of defective images.

抄録全体を表示

In this study, we focus on Nagasaki City, one of the tourist cities, and analyze how individual satisfaction among tourists affects overall satisfaction by tourist attributes and trip details in order to improve tourist satisfaction. The relationship between overall satisfaction and individual satisfaction is analyzed using correlation coefficients. We also propose a method for estimating the correlation coefficients obtained according to tourist attributes and trip details. The analysis uses data obtained from an Internet-based questionnaire survey of tourists to Nagasaki City. In the questionnaire survey, respondents were asked to respond to overall satisfaction, 11 individual satisfaction items that were considered to influence overall satisfaction, 4 tourist attributes, and 2 trip details.

抄録全体を表示

According to the "Survey on Weekday Morning Spending" conducted by Lion Corporation in 2012, 65.1% of 1,000 company employees in their 20s to 50s in the Tokyo metropolitan area answered that they "want more" or "don't have enough" time in the morning. When those who responded that they "want more time in the morning" were asked how often and how long they want to spend time in the morning, the average frequency was 3.8 days per week, and the average time was 116.1 minutes. So, in the morning, it is time-consuming and difficult to find the desired clothes among a large number of clothes hanging on hangers. To solve this problem, this paper proposes a clothing search support system that allows users to use their precious morning time for something else instead of searching for clothes. This system is created using an Arduino, a camera, and a motion unit, and is capable of performing two actions: "putting away clothes" and "looking for clothes". The system also takes pictures of the clothes and displays them on the display, making it easy to see what clothes are available and where to find them.

抄録全体を表示

In this study, we conducted musculoskeletal simulations to optimize the spring of a walking assist device using a torsion spring, and performed comparative experiments with a device using a spring of different strength. Muscle activity during walking and lower limb joint angles were measured in the experiments. Simulation results for changing the spring of the walking assist device showed that a spring stiffness of 50 N・mm/deg is optimal. The results of the comparative experiments with and without the walking aids showed improved muscle activity in all measured muscles for two subjects, including the participant who provided motion data during the simulation and another individual. Furthermore, wearing the assist device led to distinct variations in joint angles for each individual participant. Additionally, through a comparison of the participants' physical characteristics, it became evident that factors such as leg and foot length significantly influenced the performance of the device.

抄録全体を表示

This paper describes an attempt to automatically grade a Japanese C language programming exercise report by summarizing its processing explanation in three lines. We fine-tuned the natural language processing model Text-To-Text Transfer Transformer (T5) to generate a model for a three-line summary task that outputs a summary sentence from the processing explanation of the programming exercise reports. The proposed automatic grading process, grades to the description, are based on the cosine similarity between the summary sentence output by the three-line summary task and the ground truth three-line summary. The performance evaluation uses two learning models: a livedoor model that fine-tunes live-door news articles and a C-lang model that fine-tunes C language programming exercise reports. As a result of the experiment, the livedoor model, fine-tuned using training data from news articles with many common expressions in Japanese sentences, was sufficient for the three-line summary description.

抄録全体を表示

A multivariate analysis method was used to verify whether varieties could be identified from the results obtained from HPLC and other physical property data measuring instruments for cherry tomatoes. As a result, it was confirmed that three varieties could be identified using pH and lycopene content.

抄録全体を表示