Junichi Hoshino

Real-time image mosaicing from a video sequence

The paper describes a fast and robust image registration method that can be used to create a panoramic image/video from video sequences. To estimate alignment parameters for image registration, the method computes pseudo motion vectors that are rough estimates of optical flows at each selected pixel. Using the proposed method, we implemented a software system that can, with a low-cost PC, create and display panoramic images/videos in real time.

A fitness game reflecting heart rate

Lack of exercise is known to be detrimental to health, and various exercise machines that can be used indoors in a convenient manner are available. However, because expertise is required to ascertain which exercise to do and for how long, people sometimes sustain injuries by exercising excessively and find it difficult to sufficiently benefit from exercising. Therefore, fitness machines with automatic controls are being developed so that people can exercise efficiently, and exercise programs are being designed with game elements to encourage regular exercise. Monotonous exercise makes it difficult for people to experience a sense of accomplishment while exercising, and if the degree of difficulty is not set properly, people cannot perform the appropriate amount of exercise. The present study proposes a fitness game method in which heart rate is measured in real time while playing a game and the contents of the game are actively adjusted based on heart rate, thus allowing users to perform sufficient amounts of exercise and experience a sense of accomplishment. A fitness game based on boxercise movements was designed based on the present method and then evaluated.

Head-eye Animation Corresponding to a Conversation for CG Characters

CG character animations, in which CG characters engage in conversations, are widely used in multimedia contents such as videos and games. In constructing such an animation, we should consider about many factors, such as the content of the utterance and the state of conversation, which necessitates a large amount of time and labor. To deal with this problem, this paper proposes a method, which generates the head‐eye movements of the CG characters synchronized with the conversation. In this method, so as to generate a composite head‐eye movement, the view line direction is divided into the head, eye and body rotations by sharing motion mechanism dynamically adjusted according to the view line direction. And letting the two modules generating the head and the eye movements share the same conversation state, head and eye movements synchronized with the conversation are generated. Finally, we apply the proposed method to the conversation scenes, and show that natural animation can be produced easily.

Evaluation of mental stress by physiological indices derived from finger plethysmography.

BackgroundQuantitative evaluation of mental stress is important to prevent stress-related disorders. Finger plethysmography (FPG) is a simple noninvasive method to monitor peripheral circulation, and provides many physiological indices. Our purpose is to investigate how FPG-derived indices reflect on mental stress, and to clarify any association between these physiological indices and subjective indices of mental stress.MethodsThirty-one healthy women (mean age, 22 years ± 2) participated. The participants rested by sitting on a chair for 10 min. They then performed a computerized version of the Stroop color-word conflict test (CWT) for 10 min. Finally, they rested for 10 min. FPG was recorded throughout the experiment. The participants completed a brief form of the Profile of Mood States (POMS) questionnaire before and after the test. Using the FPG data, we conducted chaos analysis and fast Fourier transform analysis, and calculated chaotic attractors, the largest Lyapunov exponent, a high-frequency (HF) component, a low-to-high-frequency (LF/HF) ratio, finger pulse rate and finger pulse wave amplitude.ResultsThe HF component decreased and the LF/HF ratio increased significantly during the test (P < 0.01), while the confusion subscale of POMS increased after the test (P < 0.05). During testing, finger pulse rate significantly increased (P < 0.001), and the finger pulse wave amplitude decreased (P < 0.001). The attractor size reduced during testing and returned to a baseline level afterwards. Although the largest Lyapunov exponent showed no significant change during testing, significant negative correlation with the tension-anxiety subscale of POMS was observed at the beginning (P < 0.01). A significant negative correlation between the LF/HF ratio and two subscales was also observed in the beginning and middle of the test (P < 0.05). There were no correlations during the rest periods.ConclusionsThe physiological indices derived from FPG were changed by mental stress. Our findings indicate that FPG is one of the easiest methods to evaluate mental stress quantitatively. In particular, the largest Lyapunov exponent and the LF/HF ratio might be associated with acute mental stress. Farther examination is needed to find any association between the physiological indices and various types of mental stress.

bubble cosmos

In the field of entertainment, fantaraction is aggressively being developed these days. Fantaraction places emphasis on entertainment and art and imparts momentary surprise or pleasure. This paper proposes a bubble display system (Fig.1) that projects an image onto a real bubble containing smoke and changes the sound effect and image when the bubble bursts. This system enables a user to interactively enjoy written characters or other images projected onto a smoke-containing bubble, and a beautiful tone sounded upon bursting of a bubble, along with a different image projected onto the smoke. To detect a bubble and determine its bursting, a camera input image is divided into areas and each area is checked for a bubble. The bubble position and size are also detected, and an image is projected from a projector.

Mounting and application of bubble display system: bubble cosmos

In this paper, we propose the soap bubble display method that the image can be projected to the real soap bubbles that white smoke entered. The position and the size of soap bubbles tossed in the air are detected with the camera. By projecting the image only to the position with the projector, the soap bubble display is realized. And the image and the sound can be interactively added when the explosion of soap bubbles are judged. We mount the soap bubble display, and evaluate usefulness as entertainment that player can enjoy.

A match moving technique for merging CG and human video sequences

Merging virtual objects with human video sequences is an important technique for many applications such as special effects in movies and augmented reality. In the traditional method, the operator manually fits a 3D body model onto the human video sequence, and generates virtual objects at the current 3D body pose. However, the manual fitting is a time consuming task, and automatic registration is required. We propose a new method for merging virtual objects onto the human video sequence. First, we track the current 3D pose of the human figure by using spatio-temporal analysis and structural knowledge of the human body. Then we generate a CG object and merge it with the human figure in video. We demonstrate examples of merging virtual cloth with the video captured images.

A match moving technique for merging CG cloth and human movie sequences

Merging computer-generated objects onto a human video sequence is an important technique for many applications such as augmented reality and special effects in movies. In this paper, we propose a new method for merging virtual objects onto the human video sequence. First, we track the current 3D pose of the human body by using spatio-temporal analysis. Then we generate CG objects and merge them with the human body in video. In this paper, we demonstrate examples of merging virtual cloth with the video-captured images. Copyright

Intelligent storyboard for prototyping animation

Storyboard is a classical animation tool to help the creators to organize scenes. In this paper, we propose a new method for producing animation from storyboard. The 3D position and behavior of the characters are estimated from 2D views using the constraints optimization and examplebased interpolation method. We also show animation examples produced from the storyboard.

SanjigenJiten: computer assisted language learning system within a 3d game environment

Imagine being able to approach any object in the real world and instantly learn how to read and pronounce the name of the object in any other language. This paper proposes the use of a system that simulates this idea by utilizing the video game medium in a way that makes learning a new language simple and fun. The system was designed specifically for the new technologically-inclined generation that might benefit greatly from learning within a game environment. The process of learning a new language with this system strays from previous and conventional methods in that it employs a more visual-spatial approach to learning. Additionally, this system engages the player through the use of industry-standard video game elements such as a 3D environment, controllable main character, item collection system, scoring system, and complex rewards system. By keeping in line with what people expect from standard video games, this game is capable of holding the players attention for longer periods of time than when compared classes, textbooks, or tutors.