This study examined the effectiveness of blind spot monitoring systems using augmented reality (AR) technology. Using a VR driving simulator, 11 participants conducted experiments with three information display methods (mirror, meter panel, AR) across three driving scenarios (general roads, highways, traffic jam). Analysis of eye-tracking data and NASA-TLX subjective evaluations revealed that road conditions had a greater impact on visual attention than information display methods. AR displays reduced dangerous glance frequency by approximately 60% compared to traditional mirror displays, showing a large effect size (Cohen's$\mathrm{d} = 0. 8 1$). Traffic jam conditions showed the highest frequency of gaze deviations, while highway conditions showed the lowest. The findings indicate that consideration of usage environments is crucial in the design of driver assistance systems. The superiority of spatially-correspondent AR displays provides important guidelines for automotive HMI design prioritizing intuitive hazard perception.

Early detection and accurate classification of surface defects are essential for preserving wooden cultural heritage buildings. This study presents a comprehensive comparative evaluation of six state-of-the-art semantic segmentation methods (FCN, PSPNet, DeepLabv3+, DANet, SegFormer, Swin Transformer) for wood surface defect detection. Using a dataset of 2,000 images containing ten defect categories, we evaluated each method's performance under unified experimental conditions. SegFormer achieved the highest performance with 68.83% mIoU and 74.27% mAcc, demonstrating the effectiveness of its hierarchical encoder design for complex wood texture patterns. FCN and DeepLabv3+ achieved competitive performance with shorter training times, making them promising practical options.

Virtual reality teleportation using hand tracking faces significant challenges in vertical navigation, with conventional parabolic methods requiring users to mentally calculate trajectories and landing points. We present two novel linear pointer-based teleportation methods—SphereBackcast and Penetration—that enable intuitive vertical movement through direct pointing and collision handling strategies. Through two experiments involving 34 participants total, we evaluated these methods against traditional parabolic teleportation across diverse environments ranging from flat terrain to multi-level structures with 4m height differences. Results demonstrate that the Penetration method significantly outperforms parabolic teleportation, achieving 53% faster completion times (13.85s vs 29.38s) and 35% lower path deviation (0.552m vs 0.851m) in environments with 2m+ vertical elements. Controller input provided 12-18% performance improvements over hand tracking while maintaining consistent relative advantages of linear methods. Both proposed methods received superior usability ratings (SUS: 69.2 and 68.75 vs 64.66) and reduced cognitive workload (NASA-TLX: 25.35 and 25.08 vs 33.05), with 50% of participants preferring the Penetration method. These findings establish efficient teleportation techniques that address critical limitations in current VR navigation, offering practical solutions for applications requiring vertical movement such as adventure games and architectural visualization.

E-sports has seen significant global growth with notable economic impact, particularly in the genre of First Person Shooting (FPS) games, which are popular for their immersive play and spectator appeal. FPS games require high strategic and manipulative skills scientifically-based training methodologies are becoming crucial. Key research areas include the interaction of psychological stress with cognitive skills like reflexes, situational awareness, and reaction time. This paper explores the analysis of heart rate and visual attention in FPS games, as well as the influence of personality traits on manipulation skills. The experiment is conducted by asking participants recruited at the university to play two different shooting games and collect data. Heart rate variability is measured with an electric heart rate monitor, and an eye tracker is used to track eye gaze and examine its relationship to hit rates. The amount of mouse manipulation is recorded by the operation recording software and camera, and obtained by hand motion analysis with MediaPipe. Participants’ confidence in their shooting skills and communication intentions were surveyed using a 5 -point scale questionnaire based on the 16 Personalities framework derived from Jungian psychology. The study found that wrist-centered manipulation methods improved shooting accuracy, and players with the “ISTJ” personality type displayed higher shooting abilities.

The goal of this research is to create a reality-based experiential soccer game using XR. Users are asked to kick a ball toward a screen that has a shock-absorbing function. The trajectory of the ball at that time is projected on the screen by our proposed system. The initial velocity, axis of rotation, and number of rotations obtained from image recognition were calculated and used to calculate the trajectory of the ball and create the video to project a screen. Therefore, we succeeded in creating an environment that allows practice in a small space. User questionnaire evaluation by 97 subjects was conducted, and although the subjects with little soccer playing experience did not feel any discomfort, the more experienced the subjects were, the more they felt discomfort with the initial speed and trajectory.

A scalable codec has been constructed by using transform coding and the basic modules for scalable encoder and decoder. It allows users to choose a variety of scalable configurations in the frequency domain. The basic module is a quantizer that can quantize MDCT (modified DCT) coefficients transformed from a variety of frequency regions. This module mainly works at bit rates of more than 8 kbit/s. We can also change the target frequency regions of the basic module's input-output signals in each transform frame; i.e., we can change the scalable structure according to the nature of the input signals. In the scalable codec described here, the input-output signals are monaural and the sampling frequency is 24 kHz. The total bit rate of this scalable codec is more than 8 kbit/s. Subjective quality evaluation tests, mainly for musical sound sources, showed that it's sound quality is better than that of an MPEG-2 layer 3 codec at 8, 16, and 24 kbit/s when our scalable codec is constructed of 8-kbit/s basic modules. In combination with AAC (advanced audio coding), our scalable codec will be chosen as an international standard in ISO/IEC-MPEG-4/Audio.