부산대학교 도서관

Mobile photography guidance has emerged as a highly regarded research area within the field of human-computer interaction. With the widespread adoption of smart-phones, individuals increasingly rely on the built-in cameras of their mobile devices for photography. However, the majority of users lack knowledge on how to capture aesthetically appealing photographs, which poses a significant challenge within the realm of human-computer interaction. To assist users in enhancing their photography skills, we designed a robust and dependable aesthetic evaluation model based on a large-scale dataset collection. This model has been deployed on mobile devices to create a photography software known as “AI Photography Tutor”. This mobile application offers real-time feedback, presenting information about light, color, composition, and overall aesthetic scores to guide users in capturing more artistically appealing photos. To evaluate the performance and guidance effectiveness of this application, we conducted comparative experiments with three other photography applications employing different guidance methodologies. We performed both subjective and objective analyses of the experimental results. The findings from these experiments demonstrate that our proposed photography guidance approach, along with the mobile application developed based on it, significantly enhances the photography skills of smartphone users. Furthermore, results from user satisfaction surveys indicate that “AI Photography Tutor” surpasses other applications in terms of interactivity, usability, applicability, improvement in photography results, and overall satisfaction. We believe that this innovative application holds the potential to become an essential tool for photography enthusiasts, offering robust support for users to capture higher-quality photos.