Earticle

While generative AI's Video-to-Video technology has rapidly advanced, challenges remain in video production requiring complex camera movements and precise angle control. This study proposes a workflow that validates precise camera control by inputting reference videos created with Unity game engine into Video-to-Video generative AI. We designed five camera control missions (120⁰ rotation, vertical crane shot, compound dolly zoom, spiral orbit, and multi-angle transition) and created reference videos using Unity for each mission, then generated final videos using Runway Gen-3's Video-to-Video functionality. Focus group interviews with 5 video production experts showed that Unity references accurately conveyed camera movements across all missions (average 4.6/5 points), maintaining high accuracy (4.5 points) even in complex spiral orbit missions. Experts identified frame-level precise control, ease of iterative modification, and predictable results as key advantages, with all participants expressing intention to apply this in professional practice. This study empirically demonstrates that game engine-based reference generation is a practical solution to camera control challenges in generative AI video production.

This study employed a mixed-methods approach to examine how international students’ AI-assisted academic and career coaching needs vary according to their levels of academic and career competencies. Quantitative and qualitative analyses revealed a three-stage pattern of coaching needs, that is to say, introductory, intermediate, and implementation, across both competencies, and identified four competency-combination types: latent-vulnerable, direction-seeking, implementation-gap, and self-development. These types differed not only in the intensity of their coaching needs but also in their attitudes toward AI use, modes of expressing needs, and levels of emotional anxiety. Notably, the same AI tool functioned as a developmental resource for some groups, while for others it appeared to heighten anxiety and avoidance. The study underscores the need to view international students as adult and lifelong learners at different developmental stages and highlights the importance of competency-based, tailored AI coaching systems for effective support.

This study examines the dual potential of gaming motivations, which have been discussed as risk correlates of gaming disorder and as possible sources of vitality, by estimating wave-specific associations with Adaptive Game Use (AGU) across five waves. Five-year panel data from the Korea Creative Content Agency (Wave 1 to 5, N=778) were analyzed using MIMIC models. The results indicated that Reward and Escapism motives were consistently and positively associated with AGU across all waves. In an adaptive context, reward-oriented play may align with competence-related gratification, while escapism may reflect psychological restoration. Social motivations also showed wave-dependent patterns. Teamwork was a significant predictor in the earlier waves (Wave 1 to 3), whereas Socializing emerged as a key predictor in the later waves (Wave 4 to 5). Overall, the findings suggest that the implications of gaming motivations depend on the outcome under consideration, supporting game literacy education that strengthens the capacity to harness gaming as a positive resource rather than relying solely on restrictive approaches.

We present Text2VR, a modular pipeline that generates an interactive virtual reality (VR) scene and a runnable Unity project from a single text prompt for early-stage VR game prototyping. Given a textual description, Text2VR synthesizes a 360° panorama, decomposes it into background and foreground objects, reconstructs a 3D Gaussian Splatting background and textured 3D assets, and automatically aligns and packages them in Unity with colliders and basic interaction scripts. We evaluated the pipeline in a preliminary user study (N=14) using 7-point Likert scales on perceived ease of use, immersion, visual quality, and time acceptability. Results indicate high ease of use and immersion, adequate visual quality for rapid prototyping with limitations in complex indoor layouts, and moderate acceptance of approximately one hour of generation time when intermediate outputs are visualized. These findings suggest that Text2VR can reduce iteration costs by bridging generative AI pipelines and game-engine-ready VR content.

This paper proposes a prompt–optimization framework for generating style–consistent game images using Stable Diffusion XL. Given a reference game item image, the system first extracts an initial prompt using a vision–language captioner and a domain–specific prompt bank. The extracted prompt is converted into a list of noun-like elements, and a genetic algorithm searches for compact combinations of these elements under dual gating based on SSIM and CLIP scores. The best combinations are treated as a “style template” that can reproduce the reference image with high structural and semantic similarity. We then investigate whether this template can be reused when the main object is changed while preserving the original visual style. Experiments on fantasy-style item images show that the framework reconstructs reference images using only 8–9 automatically discovered prompt elements, and that changing the main object token together with associated detail elements yields image sets that share a consistent visual style. In contrast, naïvely replacing only the main object token often produces visually ambiguous or stylistically inconsistent images. These results demonstrate that combining automatic prompt extraction from images with evolutionary optimization provides a concrete example of style–preserving prompt design for game item image generation.

This study proposes an extended ontological model for classifying non-scripted content genres by integrating the dynamic relational framework of Actor-Network Theory (ANT) to overcome the limitations of static structural analysis inherent in existing ontology models. While previous ontology-based classification systems successfully systematized the structural attributes of non-scripted content (e.g., using 12 dimensions from the Ontojogo model), they failed to adequately capture the dynamic interactions and network formation processes among the diverse Actors surrounding the content. Drawing on Bruno Latour's ANT, which symmetrically treats human and non-human Actors in the formation of heterogeneous networks, this research re-conceptualizes non-scripted content as a dynamic network entangled with producers, cast members, viewers, platforms, and algorithms. The study integrates ANT's core concepts—Translation, Obligatory Passage Point (OPP), and Actor Dispersion—into the existing ontological framework. Specifically, the research expands six key structural attributes (Available, Game Mode, Information, Character Choice, Player Perspective, Progression) by incorporating ANT concepts, deriving new analytical dimensions such as Platform Actancy, OPP Type and Intensity, and Information Translation and Stabilization. Based on this extended model, a new genre classification system is proposed, categorized by the stabilized translation process of the network, the type and intensity of the core OPP, and the pattern of Actor dispersion.

This study examined heterogeneous longitudinal patterns of adaptive game use (AGU), maladaptive game use (MGU), and escape motivation. While prior research has emphasized pathological outcomes, intensive gaming may also serve adaptive functions. A group-based multi-trajectory model was applied to a national five-wave panel to model AGU, MGU, and escape motivation simultaneously. Results supported five co-development profiles: Low-Risk Decreasing (G1); Low-Risk Mixed Increasing (G2), with rising AGU and MGU and slight increases in escape motivation at low levels; Moderate-Risk Recovery (G3), with stable AGU but declining MGU and escape; High-Risk Deteriorating (G4), with increasing AGU and escape and a curvilinear escalation in MGU; and Chronic High-Risk (G5), with persistently elevated MGU and escape across waves. Overall, adaptive and maladaptive aspects of game use may coexist and codevelop rather than forming a single continuum, underscoring the need to distinguish high engagement from risk trajectories and to tailor interventions to motivational processes.

This study derives design principles for integrating serious game elements into bidirectional exercise guidance systems in smart senior centers. South Korea's rapidly aging population, particularly in rural areas, faces limited healthcare access and social isolation. The Smart Senior Center initiative uses ICT to connect instructors with multiple centers, yet sustaining elderly participation remains challenging. Through qualitative interviews with expert groups (practitioners, health specialists, HCI experts) and elderly participants at Ganghwa County Smart Senior Center (November-December 2024), five themes emerged: regional identity-based community formation, visual accessibility requirements, immediate feedback importance, balanced competition-collaboration design, and critical support staff roles. Seven design principles were derived: reflecting cultural context, prioritizing visual accessibility, providing immediate multi-sensory feedback, facilitating group interaction, balancing competition and collaboration, establishing human support systems, and ensuring technical simplicity. This research provides practical guidelines for elderly digital healthcare interaction design, contributing to health promotion and quality of life improvement. this study proposes a new analytical perspective for understanding digital image environments.

Contemporary virtual reality(VR) media is expanding beyond mere reproductions of the physical world into an independent artistic medium that creates novel sensory experiences. This study aims to investigate the complex mechanisms through which surreal virtual landscapes paradoxically generate a heightened sense of sensory realism. Drawing on theories of embodied presence in VR and the duality of perception, the analysis focuses on how the interaction between organic objects and unrealistic light phenomena affects audience perception. The results confirmed that surreal representations act as a core mechanism that heightens sensory immersion by inducing cognitive tension in the user. The autonomous movement of virtual objects imparts a sense of otherness and presence, while multisensory elements such as auditory reverberation contribute to the perceived physicality of the virtual space. Furthermore, interactive compositions that respond to user movement reinforce a sense of agency, transforming the virtual environment into an active experiential arena. This study demonstrates that virtual space serves as a creative domain where the sense of reality and the imagery of the virtual merge to generate a new perceptual dimension. Surreal elements do not diminish realism; rather, they awaken perception and offer renewed insights into space and presence. The findings reframe VR media not merely as technological progress, but as an artistic tool that expands human sensory experience.