프로젝터-카메라 시스템의 객체 감지 향상을 위한 Unet 기반 이미지 복원

초록

The projector-camera system is a technology widely used in fields such as augmented reality (AR), education, and healthcare, but it suffers from degraded object recognition performance due to color distortion caused by the projector light. To address this issue, this study proposes a U-Net-based color restoration model. The proposed model takes color-distorted images as input and restores them to colors close to the original, which was experimentally validated through improvements in image similarity and object recognition accuracy. In addition, the model operates independently from object recognition systems, allowing flexible application under varying lighting conditions without retraining. The results demonstrate the model’s effectiveness in resolving color distortion issues and improving recognition performance.

키워드