Earticle

초록

영어

This paper presents a method to optimize 3D Gaussian Splatting, improving computational efficiency for real-time applications on edge devices like Jetson AGX Xavier. By applying pruning and quantization techniques, we enhance computing speed with minimal degradation on image quality. Our method enables efficient 3D scene reconstruction on resource-limited devices, making it suitable for AR/VR and autonomous driving.

목차

Abstract
I. INTRODUCTION
II. RELATED WORK
A. 3D Reconstruction
B. Network Pruning and Quantizaiton
III. METHOD
A. Overview
B. Gaussian Clustering with HDBSCAN
C. Gaussian Pruning
IV. EXPERIMENTS
A. Experimental Settings
B. Evaluation of Pruning Ratio
C. Performance Comparison on Edge Device
D. Ablation Study
V. CONCLUSION
ACKNOWLEDGMENT
REFERENCES

저자

• Kang-San Kim [ Department of Intelligent Systems and Robotics Chungbuk National University Cheongju, Republic of Korea ]
• Jae-Hyeok Kang [ Department of Intelligent Systems and Robotics Chungbuk National University Cheongju, Republic of Korea ]
• Yeong-Ju Lee [ Department of Intelligent Systems and Robotics Chungbuk National University Cheongju, Republic of Korea ]
• Youngbae Hwang [ Department of Intelligent Systems and Robotics Chungbuk National University Cheongju, Republic of Korea ] Corresponding Author

참고문헌