Paper
An Empirical Study of User Playback Interactions and Engagement in Mobile Video Viewing
2025. IEEE AccessEunnho Kim, Seungjae Oh, Sangkeun Park
Mobile video viewing on popular platforms such as YouTube and Netflix is widespread, yet the role of specific viewing interactions in shaping user engagement remains underexamined. This study investigates how skipping behaviors (including their types and directions) and playback speed adjustments relate to user engagement, with a focus on video abandonment and user satisfaction. We developed a custom mobile web application for video viewing and collected viewing logs and self-reports from 25 participants during two 10-day field studies. Our findings reveal that different skip types and directions are associated with distinct engagement outcomes. For example, scrubbing often correlates with higher abandonment, whereas backward skips may indicate greater engagement. Playback speed adjustments can signify deeper involvement, allowing users to tailor their viewing speeds without missing key content. Notably, video abandonment did not always equate with dissatisfaction; some users left after meeting their immediate viewing goals. These insights suggest that users’ playback interactions may serve as indicators of user engagement and can be incorporated into video recommendation systems to enhance user satisfaction. We conclude by discussing the design implications of enhancing user satisfaction.
가속도 센서를 이용한 CNN 기반의 스마트폰 터치 에러 감소 기법 개발
2024. Korea Conference on Software Engineering 2024 (KCSE2024)김은호, 박상근
스마트폰의 보급률이 높아지면서 사람들은 일상생활에서 여러 모바일 서비스를 편리하게 이용할 수 있게 되었다. 스마트폰 사용자는 정지된 상황뿐 아니라 이동 중에도 한 손으로 스마트폰을 조작하며 모바일 서비스를 이용하기도 한다. 이동 중에 스마트폰을 사용하면 사용자의 신체 흔들림 등으로 사용자의 한 손 터치 정확도가 떨어지는 불편함이 있다. 본 논문에서는 가속도 센서를 통해 측정되는 스마트폰의 기울어진 정도를 기반으로 사용자가 터치하려는 버튼을 예측하는 CNN 모델을 개발했다. 또한 예측된 버튼의 터치 허용 영역을 유연하게 확장하는 터치 에러 감소 기법을 제시하고 실험을 통해 그 유용성에 대하여 검증했다.