Visualization of agreement and discussion processes during computer-supported collaborative learning

This study examined the effects of the shared space (SS) on students’ behaviors in a computer-supported collaborative learning (CSCL) environment. The SS visualizes discussion and agreement during online discussions. It was hypothesized the SS would increase the media richness of the CSCL-environment, would stimulate critical and exploratory group-norms, would lead to more positive perceptions of online collaboration, and would have an impact on students’ collaborative activities. In total, 59 students working in 20 groups had access to the SS visualization, while 58 students working in 20 groups did not. The results show that students with access to the SS visualization: (a) perceived higher media richness; (b) had a more exploratory group-norm perception; (b) perceived more positive group behavior; (c) perceived their group’s task strategies to be more effective; (d) engaged in different collaborative activities and (e) performed better on one part of the group task. These results demonstrate the potential benefits of visualizing agreement and discussion during CSCL.     

Context enabled decision aids to support the situation awareness and performance of risk engineers carrying out loss prevention surveys

Routine inspection by insurance companies at their clients’ facility, also known as loss prevention survey, help identify the best strategies to minimize damages when there is a high-speed wind event. More specifically, wind vulnerabilities associated with a building are evaluated using a process known as windstorm risk inspection. This routine inspection helps clients reduce the extent of damages caused by high-speed wind events including hurricane and tornado. Risk engineers make use of their subjective and analytical deduction skills to successfully carry out the inspection tasks. In this research the researchers investigated the effect of context-based visualization strategies on situation awareness and their understanding of the situation. The study examined how different types of information contribute towards the three levels of situation awareness. Following a between-subjects study design, 65 participants completed the study. Each session lasted 90–120 min. A checklist based and predictive display-based decision aids were tested and found to be effective in supporting the situation awareness requirements as well as performance of risk engineers. However, the predictive display only helped with certain tasks such as understanding the interaction among different components on the rooftop. For remaining tasks such as perceiving obvious issues like membrane tear, clogged drains and vegetation growth, checklist alone was sufficient. This study helped the understanding of the advantages and disadvantages of the decision aids tested. More specifically, these decision aids can improve the mental model of novice risk engineers. Additionally, this study provided insights that could help design training materials for infrastructure inspectors.     

Lessons learned: Evaluating visualizations for occluded objects in handheld augmented reality

Handheld devices like smartphones and tablets have emerged as one of the most promising platforms for Augmented Reality (AR). The increased usage of these portable handheld devices has enabled handheld AR applications to reach the end-users; hence, it is timely and important to seriously consider the user experience of such applications. AR visualizations for occluded objects enable an observer to look through objects. AR visualizations have been predominantly evaluated using Head-Worn Displays (HWDs), handheld devices have rarely been used. However, unless we gain a better understanding of the perceptual and cognitive effects of handheld AR systems, effective interfaces for handheld devices cannot be designed. Similarly, human perception of AR systems in outdoor environments, which provide a higher degree of variation than indoor environments, has only been insufficiently explored.In this paper, we present insights acquired from five experiments we performed using handheld devices in outdoor locations. We provide design recommendations for handheld AR systems equipped with visualizations for occluded objects. Our key conclusions are the following: (1) Use of visualizations for occluded objects improves the depth perception of occluded objects akin to non-occluded objects. (2) To support different scenarios, handheld AR systems should provide multiple visualizations for occluded objects to complement each other. (3) Visual clutter in AR visualizations reduces the visibility of occluded objects and deteriorates depth judgment; depth judgment can be improved by providing clear visibility of the occluded objects. (4) Similar to virtual reality interfaces, both egocentric and exocentric distances are underestimated in handheld AR. (5) Depth perception will improve if handheld AR systems can dynamically adapt their geometric field of view (GFOV) to match the display field of view (DFOV). (6) Large handheld displays are hard to carry and use; however, they enable users to better grasp the depth of multiple graphical objects that are presented simultaneously.     

Master-CAM: Multi-scale fusion guided by Master map for high-quality class activation maps

Class Activation Map (CAM) is one of the most popular approaches to visually explain the convolutional neural networks (CNNs). To obtain fine-grained saliency maps, some works fuse saliency signals of the same image at larger scales. However, existing methods based on multi-scale fusion cannot effectively remove the noise from larger-scale images. In this paper, we propose Master-CAM, which uses Master map to guide multi-scale fusion process to obtain a high-quality class activation map. Master-CAM utilizes the general localization ability of the Master map to reduce the noise of the maps. We call the one with the general localization ability among the saliency maps from the same image as Master map, which is the saliency map of the original-scale input in the multi-scale scenario. In addition, we also present a simple yet effective fusion strategy, Master-Fusion, which is derived from the fusion operation in Master-CAM. Master-Fusion strategy can be easily attached to some saliency methods to improve the performance of these methods. We show through qualitative and quantitative experiments that the proposed Master-CAM outperforms the state-of-the-art methods in different CNN frameworks and datasets.