Publications
JOURNAL (INTERNATIONAL) Investigation of Smartphone Grasping Posture Detection Method Using Corneal Reflection Images Through a Crowdsourced Experiment
Kaori Ikematsu*, Xiang Zhang* (Keio University), Kunihiro Kato (Tokyo University of Technology), Yuta Sugiura (Keio University), *co-1st authors
International Journal of Human–Computer Interaction (IJHCI)
June 25, 2025
Understanding smartphone users' grasp postures can improve the mobile interaction experience; for instance, the user interface (UI) layout can be dynamically adjusted based on the detected grasp. In our previous work, we introduced ReflecTouch, a grasp posture detection method that leverages corneal reflection images captured by the front camera. Building on this, we conducted a follow-up study using crowdsourced experiments in uncontrolled, in-the-wild environments. Based on the results, we discuss design considerations and practical strategies, including recommendations for integrating ReflecTouch with complementary sensing modalities, such as touch and IMU-based methods, to enhance the model's accuracy in real-world settings.
Paper :
Investigation of Smartphone Grasping Posture Detection Method Using Corneal Reflection Images Through a Crowdsourced Experiment
(external link)