In this paper, we present a prototype Virtual Welding Simulator, which supports interactive training of welding process by using multimodal interface that can deliver realistic experiences. The goal of this research is to overcome difficult problems by using virtual reality technology in training tasks where welding is treated as a principal manufacturing process. The system design and implementation technical issues are described including real-time simulation and visualization of welding bead, providing realistic experience through 3D multimodal interaction, presenting visual training guides for novice workers, and visual and interactive interface for training assessments. According to the results from an initial user study, the prototype VR based simulator system appears to be helpful for training welding tasks, especially in terms of providing visual training guides and instant training assessment.

In this paper, we present a virtual training system which realistically represents the situation of real welding. First of all, we built a database about welding outputs such as the shape of bead which is the deposit outcome resulting from inputs of real welding conditions and operations. Second, we performed an analysis of relations between input variables and output variables, and the major and minor factors influencing on the welding shape were classified. Finally, we designed an estimation process of outputs by various inputs such as user's movement, and constructed the method of graphical representation for real-time visualization from heuristic sources. Additionally, we also installed a welding simulator for a teacher and trainees, which can support not only a variety of welding situations but prompt evaluation of results and educational guide by optimal welding conditions.

In this paper, we propose a new immersive virtual aquarium system that comprises enhanced navigation, shared user experience, and manipulation in virtual reality. The system utilizes a large studio that allows users to walk around naturally and operates as a surrogate for navigation in the virtual aquarium. It also provides an additional view that shows what is happening in the virtual aquarium to observers. The system can be regarded as a form of augmented virtuality as it superimposes a live video avatar of users into the virtual aquarium. In addition, each user can interact with virtual fishes via 3D gestures because the system also utilizes a gesture recognition method that is suitable for the physically large space. These features enable our proposed system to give users and observers a more immersive and novel experience than conventional virtual reality systems.

This paper presents a study on how augmenting semi-synthetic image data improves the performance of human detection algorithms. In the field of object detection, securing a high-quality data set plays the most important role in training deep learning algorithms. Recently, the acquisition of real image data has become time consuming and expensive; therefore, research using synthesized data has been conducted. Synthetic data haves the advantage of being able to generate a vast amount of data and accurately label it. However, the utility of synthetic data in human detection has not yet been demonstrated. Therefore, we use You Only Look Once (YOLO), the object detection algorithm most commonly used, to experimentally analyze the effect of synthetic data augmentation on human detection performance. As a result of training YOLO using the Penn-Fudan dataset, it was shown that the YOLO network model trained on a dataset augmented with synthetic data provided high-performance results in terms of the Precision-Recall Curve and F1-Confidence Curve.

No abstract