[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This study aimed to develop a method for detecting difficult-to-read characters by tracking the gazes of personal computer (PC) users. A method for detecting English words that are difficult to read was proposed previously. However, this method required a dedicated user interface because the accuracy of the eye tracker is low (eye-tracking errors). In actual PC work, the range of use of a dedicated user interface is limited, and efficient work may be hindered. In this study, we propose a character detection method with eye-tracking error correction using projective transformation. The usefulness of the proposed method was examined based on the detection accuracy. The results showed that the proposed method was effective in detecting difficult-to-read characters in F-measure > 0.58 in the first row of the text. Moreover, the usefulness of the proposed method was demonstrated for nine rows or shorter texts compared with no-correction eye-tracking data.

View full abstract

The use of non-contact information to estimate the psychological state is becoming popular as a new technology to support communication, particularly with the development of information devices and the impact of COVID-19. However, facial expressions and emotions are influenced by culture, situations, and individual differences. Therefore, it is difficult to estimate the emotional arousal interval using facial expressions alone. Hence, it is necessary to include physiological changes such as facial skin temperature. Previous studies have shown that changes in the skin temperature occur in the nose, right cheek, and left cheek during positive emotional arousal. Moreover, for estimating emotional arousal using thermal and visible images of faces, the facial thermal images evidently deliver better results than the visible images. However, for practical use, it is necessary to focus on time-series changes and conventional methods have not yet been developed. In addition, results combining thermal and visible images have not been presented to date. In this study, we propose a method for estimating the intervals between aroused and unaroused emotions using long short-term memory and chronological skin temperature changes in the nose, right cheek, left cheek, and mouth angle.

View full abstract

[in Japanese]

View full abstract

Completely automated public Turing test to tell computers and humans apart (CAPTCHA) distinguishes between human and machine access to web pages. However, CAPTCHAs that use visual characteristics have been studied to make machine responses more difficult. In this study, a model that reproduces human color perception was constructed using a neural network with the aim of utilizing the ambiguity of human color perception in CAPTCHA. The results of the investigation of training data showed that the data used to train the constructed color perception model was optimal when only colors that could be classified as categorical colors were used.

View full abstract

[in Japanese]

View full abstract

In recent years, the number of patients with diabetes mellitus has increased. Moreover, the prevalence of diabetes among older age groups in Japan is notably higher, indicating a growing proportion of patients with diabetes. Diabetic foot complications manifests in some patients with diabetes mellitus; hence, timely identification of the symptoms associated with diabetic feet is crucial for preventing severe complications. It is imperative for patients to observe their feet regularly; however, recognizing diabetic foot symptoms can be challenging for individuals without medical expertise owing to the variability of such symptoms. In this study, we focused on tinea unguium as a case type and utilized machine learning to classify images of tinea unguium and normal feet. The evaluation results showed that the combination of ResNet-50 and a support vector machine yielded the best performance when applied to a dataset of the acquired images of the nail regions the feet.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

The trial production of a new concept quadcopter with a vertical-axis wind turbine (VAWT) is presented for power-saving during forward flight. Continuous transition from hovering to power-saving level flight can be done only by tilting forward of the quadcopter. The VAWT rotates with the downwash of its rotors and against the wind by forward flight, and the lift is generated. The lift of the VAWT during level flight partially supports the weight of the aircraft, and it causes the reduction of power consumption and extend the flight area.

View full abstract

[in Japanese]

View full abstract

In this study, I will create security educational materials targeting robot control. It's because robot control is easy to interest mechanical and electrical engineering students and engineers, as well as junior high and high school students with no expertise in the field. Therefore, I created a virtual environment that mimics an actual robot control environment. I will launch attacks against the virtual environment and create educational materials based on what I learn. In this study, a pseudo-environment was constructed with "RoverC.Pro", a robot controllable by ROS(Robot Operating System). ROS built on Docker. The program on the RoverC.Pro side is created by Arduino and performs simple operations depending on what it receives. The ROS uses "Topic" as its communication method. In ROS, the contents of communications are exchanged unencrypted. I took advantage of this to launch a man-in-the-middle attack.Since we were able to confirm that this attack could tamper with the contents of communications and threaten to halt communications, this paper describes what we did leading up to the creation of the educational material.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Waste-printed circuit boards (WPCBs) are recycled in nonferrous metal smelters to recover valuable metals. Workers visually determine the similarity between images of previously handled WPCBs and images of substrates to be measured for their valuable metal content and classify the substrates to so that appropriate processing methods can be selected. In particular, groups of substrates of multiple types (mixed WPCBs) are more difficult for workers to process. Therefore, in this study, we propose a method to evaluate different types of substrates using segmented images of mixed WPCBs. Our results show that the proposed approach was able to classify substrates appearing in segmented images well.

View full abstract

With the recent emergence of a super-aging society, extending the healthy life expectancy of older adults has become increasingly important. Dementia has been reported to hinder the extension of healthy life expectancy. Electronic sports (eSports) have recently attracted wide attention to prevent dementia. People experience various emotions while playing eSports. The quantification of these emotions indicates the extent to which they enjoy playing eSports. This indicator can be used to evaluate eSports activities. We propose a method to estimate when emotions occur while playing eSports with the aim of creating a quantitative index to determine people’s emotions while playing eSports. Consequently, we determined the possibility of estimating emotional arousal intervals by examining the changes in the saturation of the cheeks using machine learning.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract