In education for children and/or guidance of sports, it is important to give suitable instruction to children or players, whom we called agents in this paper. It is necessary to understand the ability and characteristic of the agents by observing the learning processes and to change the teaching methods as needed. In this paper, we consider the learning parameter estimation and adequate rewarding method, using simulation which agents learn maze problem. For learning of the maze, we used Q-learning well known in the field of reinforcement learning. And we conducted experiments using multiple agents with different learning parameters. Agents’ action data at the early stage of learning is used for learning parameter estimation using self-organizing map. After that, we change rewarding method based on the estimated agents’ learning parameters.

View full abstract

For the people who lost the walking ability temporally due to traffic accidents or falling, the timely walking rehabilitation training is necessary significantly. In the former study, an intelligent walking training robot (IWTR) was developed, and its effectiveness was verified by clinical testing for walking rehabilitation, in which the user was instructed to follow a series of designed walking paths. However, as the walking ability recovery, it is demanded more free walking training with the insurance of walking safety. In this study, we proposed a walking intention recognition method based on the visual information of patients’ feet, by which IWTR can recognize the patients’ walking intention for tracking them to provide walking assistance. Therefore, IWTR can keep patients from the falling risk during their spontaneous walking training and the patients can walk on plane field without moving limitation. In the proposed system, the patients’ feet motion was monitored by a camera fixed on the IWTR. Depending on the markers set on patients’feet, the orientation and position of each foot were extracted for walking intention recognition. By assessing the relationship between feet pattern and walking intention, eight rules were derived, and a rule-based waking intention method was proposed. We verified the walking intention recognition accuracy of the proposed method on the IWTR by experiments, and the effectiveness was presented in this paper.

View full abstract

In this study, the effects of acoustic stimuli on the heart rate fluctuations (HRFs) and autonomic nervous system activity in prenatal chick embryos was investigated. Here, instantaneous heart rate measurements were collected while the embryos were exposed to three types of acoustic stimuli – classical music, rock music, and hatchling cheeps – for 30 min per day from embryonic days 16 through 20 prior to hatching day. The results showed that the frequency of the HRFs was higher when rock music was played than when classical music was played. However, the autonomic nervous system activity did not change significantly with rock versus classical music or over the days of incubation. During the perinatal period, the effects of acoustic stimuli on the autonomic nervous system activity did not exceed the effects of physiological and biological development. This may be because the 1/f frequency component of the music is filtered out by the egg. However, the 1/f spectrum of heartbeat rhythm is not affected by the sound-filtering effect of the eggs, so it is possible that the presence of 1/f fluctuations in the maternal heartbeat rhythm affects the autonomic nervous system activity of the chick embryo. Therefore, further studies are needed to investigate the response of the embryo to the vibrational stimulus of the direct heartbeat.

View full abstract

In recent years, research in image processing is very popular and demanding because of its diverse and humongous application. Face detection i one of the dominant mechanisms for security purpose, video surveillance, humane computer interaction, etc. Numerous algorithms have been proposed by many types of research regarding face detection and Viola-Jones algorithm is one of them. This paper deals with Viola-Jones algorithm for face detection in various positions and surroundings which include distance from the camera, background color, the angle of the object, etc. Further, this paper also performs an analysis in order to find the best settings that are appropriate for ViolaJones algorithm.

View full abstract

The recognition of human actions is one of the most important areas in computer vision. It is a very difficult research area due to complexity of understanding human actions from video. In this paper, we study on the recognition of human actions based on modified Motion History Image (modified-MHI). The MHI is an extracted image from the video of human actions, which contains the necessary motion information. We proposed a way to better represent MHI, which is immune to noises due to camera movement and motion of unwanted objects. Edge detection of moving objects and finding dominant motion portion are important factors for efficient human action recognition. Finding dominant motion portion is a very challenging task because of the change of background intensity at daylight and movement of camera. We have obtained better MHIs from edge detection. The obtained MHI templates are processed afterwards to find the binary patterns and motion gra- dient features. These features are exploited for action recognition by using Support Vector Machine (SVM). We have obtained 98.1% accuracy in classifying actions on Pedestrian Action Dataset which proves the robustness of our proposed method. This dataset is comprised of 8 different actions and each action is performed by 20 different subjects.

View full abstract