This study models the decision-making characteristics of a driver regarding whether he accepts a merging car at a highway junction. Then, the application of the modeling to the design of merging behavior control is proposed. First, the driving behavior on the main lane at a highway junction is observed using a driving simulator, particularly focusing on the driver's state of decision (SOD), which represents the acceptance for merging a car coming from the merging lane. Second, the driver's SOD is modeled using a logistic regression model and the prediction performance of the identified model is verified. Finally, the speed controller of the merging car is designed to maximize the acceptance from the cars on the main lane. The key idea here is to minimize the entropy of the SOD of the driver on the main lane by optimizing the speed of a merging vehicle. This problem is quantitatively formulated using an identified decision-making model and addressed by applying a randomized approach to the optimization. This enables the automated vehicle to realize a considerate merging behavior at a highway junction. Numerical experiments are performed to demonstrate the usefulness of the proposed design scheme.

As a method to judge the benign or malignant polyp from endoscope images, some methods have been proposed using an ultra-high magnification endoscope. The ultra-high magnification endoscope enables the diagnosis at the cell level. However, it tends to spend many times for diagnosis and requires specific expensive devices. There are three types of images that are taken for diagnosis by the regular endoscope: white light, dye, and narrowband image (NBI) in general. This paper proposes a benign/malignant polyp classification method using these images taken by the regular endoscope. Each image features derived from endoscope images are extracted by adapting a pre-trained CNN to each domain. Finally, polyps are classified using extracted features. Experiments confirmed that the proposed method enabled the classification of benign or malignant colorectal polyps with over 90% accuracy.

The service presented in this paper searches two digital archives of handwritten character pattern images on Japanese historical documents. Any binary character pattern image consisting of black character shapes and white background becomes a search key. The system owns a web-based user-interface, dictionaries of character shape feature vectors extracted from the archives, and character shape similarity evaluation modules, and runs independently from metadata formats of the archives. The service displays a list of similar character pattern image for each of the archives. The images in the lists link to the archives that provide more information of the character patterns. We present the technical details and improvements of the service. Since character shapes on historical documents contain noises, we have employed a noise absorber process for the shape evaluation. The service has received more than 30,000 search requests for a month since it was introduced to the public. For each key, processing time of the shape evaluation is less than one second, and other processes including security checking need another several seconds. As an improvement, we have updated the user-interface for better usability of smartphones. Also, we have created and distributed an application for smartphones to support creating the binary image keys. Users of about 2,900 Apple IDs had downloaded the application. We introduce the latest version of the application that we will soon distribute at the Apple App Store. Evaluating the service and analyzing the search requests are our future tasks.

This paper presents a new control method for merging task at highway junction by using the model predictive control in which the decision entropy of the drivers on the main lane is explicitly considered as the cost function. Authors have already proposed the evaluation measure for the acceptance of the drivers on the main lane (supposed to be manual cars) to the merging car approaching from the merging lane (supposed to be automated car). In addition, the decision entropy of the driver on the main lane has been formally defined by using the stochastic model of the decision making. Based on this previous study, a new control method for the merging task of the automated car is addressed. The control problem is formulated so as to find the optimal speed of the merging car which minimizes the decision entropy of the drivers on the main lane. The proposed control strategy achieves the harmonized merging task in a sense that the drivers on the main lane can easily decide whether to accept or reject the cut-in of the merging car. The model predictive control is formulated as a nonlinear optimization problem, and solved by using the randomized approach. Finally, the validity of the proposed method is verified through some simulation studies.

The spatial distribution of large mammals has been increasing in industrial nations, often leading to conflicts with humans and significant alteration of natural ecosystems. Species distribution models (SDMs) of the sika deer, Japanese serow, wild boar, Japanese macaque and Asiatic black bear in Japan were constructed to identify the important factors for habitat management. We also projected two scenarios of their future range expansions under a scenario of no change in the variables, and one where the variables change according to a time series. Modelling showed that all species preferred forest and forest/agricultural ecotones. In the best models of the sika deer and wild boar, grasslands (GRASSs) were an important positive factor, while human population density and hunting pressures were negative factors. High human population density was also a negative factor with the Japanese serow and Asiatic black bear. Key environments for the distribution and expansion of large mammals were agricultural areas near forest and GRASSs. Therefore, limiting the access of large mammals to such habitats may contribute to wildlife management. In both scenarios, our projections suggest that the distributions of all species could expand by 2028, resulting in more conflict with humans and significant alterations to natural ecosystems in the future. In order to stop this damaging expansion, habitat management and landscape redesign are important. However, a SDM and future projection include uncertainty. It is necessary to validate the estimates and projections by monitoring the future changes in distribution.

We propose a method which performs an isotropic morphological dilation via implicit smoothing for the purpose of restoring the degraded character shapes of binarized images. Exploiting the idea of geodesic morphology that the binary image and its distance transformed image are interconvertible, we apply a smoothing method not to the binary image but to the distance transformed image, and then reconvert it by binarization. This allows us to apply conventional smoothing methods for continuous intensity, i.e., gray scale, images to the discrete intensity, i.e., binary, image implicitly. For instance, by using an isotropic diffusion together with geodesic dilation, an isotropic dilation along the stroke direction is obtained and brings better results.

A table look-up model of thin-film transistors has been developed for circuit simulation. This model utilizes three schemes. First, the spline interpolation with transformation by y = x+log(x) achieves precision for both on-current and off-current simultaneously. Second, the square polynomial supplement solves an anomaly near the points where drain voltage equals zero. Third, the linear extrapolation achieves continuities of the current and its derivatives as a function of voltages for areas outside where the spline interpolation is performed and improves convergence during circuit simulation.

Describes the result of a fundamental study on pattern recognition using autoassociative neural networks, and experimental comparison on handwritten numeral recognition by conventional multi-layered neural network and statistical classification techniques. As the statistical classification techniques, the projection distance method and the nearest neighbor method are employed. The relationship between the projection distance method which is based on the K-L expansion and three layered autoassociative networks is discussed, and it is shown that the three and five layered autoassociative networks are superior to the projection distance method. In the handwritten numeral recognition experiment, a total of 44862 numeral samples collected by IPTP are used to evaluate and compare the recognition rates of the autoassociative networks, the mutual associative network, the nearest neighbor method, and the projection distance method. The five layered autoassociative networks achieved the highest recognition rate in the handwritten numeral recognition experiment. The result of experiment together with the fundamental study show that the autoassociative networks have such characteristics that: (1) class independent training makes the possibility of local convergence less than that of the mutual associative network, (2) the networks possess the higher ability of dimension reduction and interpolation than the nearest neighbor method (3) they yield less misclassification due to subspace sharing than the projection method, (4) the five layered autoassociative network can fit a curved hypersurface to a distribution of patterns.