상세정보
학술대회
Lightweight Deep Learning based Intelligent Mobile Augmented Reality
- 발행일
- 202111
- 출처
- International Conference on Consumer Electronics (ICCE) 2021 : Asia, pp.373-375
- 협약과제
- 21HH5100, 객체추출 및 실-가상 정합 지원 모바일 AR 기술 개발, 고종국
- 초록
- Recently augmented reality technology is developing from simply displaying virtual objects in the real world to technologies that realistically fuse with real information. This paper introduces intelligent mobile AR technology that provides object and scene information by utilizing lightweight object detection and scene segmentation techniques operable in mobile phones. We provide object detection and scene segmentation capabilities with effective model structure improvements and model compression-based deep learning models for lightweight deep-learning models.In addition, we have made intelligent demonstration contents based on ARCore to provide object and scene information as well as display simple virtual objects.
- KSP 제안 키워드
- Augmented Reality technologies, Augmented reality(AR), Effective model, Mobile AR, Model compression, Model structure, Object detection, Real-world, Scene Segmentation, deep learning(DL), deep learning models