Chin-Yang Lin

Chin-Yang (John) Lin | 林晉暘

I am a Ph.D. student in Computer Science at National Yang Ming Chiao Tung University (NYCU), advised by Prof. Yu-Lun Liu and Prof. Wei-Chen Chiu. I am also working as an imaging engineer intern at Logitech, where I focus on GPU image quality enhancement.

Previously, I received my Bachelor's degree from National Cheng Kung University (NCKU), where I double-majored in Computer Science and Geomatics. During this time, I also worked as a Research Assistant at the HDMap Center under the supervision of Prof. Kai-Wei Chiang.

Email / CV / Google Scholar / LinkedIn / Github / YouTube

Follow @linjohnss

NYCU

Logitech

HDMap Center

NCKU

News

Aug, 2025	Start my Ph.D. journey at NYCU
Jul, 2025	One paper "See, Point, Fly" is accepted at CoRL 2025
Jun, 2025	One paper "LongSplat" is accepted at ICCV 2025
Mar, 2025	Two papers "FrugalNeRF" and "AuraFusion360" are accepted at CVPR 2025
Dec, 2024	Serve as a student volunteer for SIGGRAPH Asia 2024
Oct, 2024	Serve as a reviewer for ICLR 2025
Jul, 2024	Start my imaging engineer internship at Logitech
Mar, 2024	One paper "BoostMVSNeRFs" is accepted at SIGGRAPH 2024

Research

My research lies at the intersection of computer vision, machine learning, and 3D geometry. I focus on enabling machines to perceive and reconstruct the 3D world with efficiency, accuracy, robustness, and scalability, using learning-based methods that improve with data and computation. I am passionate about building real-time 3D reconstruction systems that generalize effectively in the real world.

	LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videos Chin-Yang Lin, Cheng Sun, Fu-En Yang, Min-Hung Chen, Yen-Yu Lin, Yu-Lun Liu ICCV, 2025 project page / arXiv / video / code Star LongSplat reconstructs long, casually captured videos into coherent compact 3D scenes by leveraging incremental joint pose and 3DGS optimization, accurate pose estimation, and efficient anchor formation.
	See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation Chih Yao Hu, Yang-Sen Lin, Yuna Lee, Chih-Hai Su, Jie-Ying Lee, Shr-Ruei Tsai, Chin-Yang Lin, Kuan-Wen Chen, Tsung-Wei Ke, Yu-Lun Liu CoRL, 2025 project page / arXiv / video / code See, Point, Fly (SPF) enables UAVs to navigate to any goal based on free-form natural language instructions in any environment, without task-specific training.
	FrugalNeRF: Fast Convergence for Extreme Few-shot Novel View Synthesis without Learned Priors Chin-Yang Lin, Chung-Ho Wu, Chang-Han Yeh, Shih-Han Yen, Cheng Sun, Yu-Lun Liu CVPR, 2025 project page / arXiv / video / code FrugalNeRF turns just two images and 10 minutes into high-quality 3D scenes by using weight-sharing voxels and cross-scale geometric adaptation to guide training without relying on external priors.
	AuraFusion360: Augmented Unseen Region Alignment for Reference-based 360° Unbounded Scene Inpainting Chung-Ho Wu, Yang-Jung Chen, Ying-Huan Chen, Jie-Ying Lee, Bo-Hsu Ke, Chun-Wei Tuan Mu, Yi-Chuan Huang, Chin-Yang Lin, Min-Hung Chen, Yen-Yu Lin, Yu-Lun Liu CVPR, 2025 project page / arXiv / video / code A reference-guided 3D inpainting approach utilizing SDEdit on aligned Gaussian initialization, and created a 360° inpainting dataset (360-USID) for comprehensive evaluation.
	DiffIR2VR: Zero-Shot Video Restoration with Diffusion-based Image Restoration Models Changhan Yeh, Chin-Yang Lin, Zhixiang Wang, Chi-Wei Hsiao, Ting-Hsuan Chen, Yu-Lun Liu arXiv, 2024 project page / arXiv / code / demo DiffIR2VR-Zero leverages pre-trained diffusion models for video restoration, using hierarchical token merging and hybrid optical flow with nearest neighbor matching. It achieves top performance across diverse degradations without training.
	BoostMVSNeRFs: Boosting MVS-based NeRFs to Generalizable View Synthesis in Large-scale Scenes Chih-Hai Su, Chih-Yao Hu, Shr-Ruei Tsai, Jie-Ying Lee, Chin-Yang Lin, Yu-Lun Liu SIGGRAPH, 2024 project page / arXiv / video / code Our method does not require training and can adapt to any MVS-based NeRF methods in a feed-forward fashion to improve rendering quality.
	A CNN-Speed-Based GNSS/PDR Integrated System for Smartwatch Chin-Yang Lin, Yang-En Lu, Chi-Hsin Huan, Kai-Wei Chiang MMT, 2023 paper / video We proposed a GNSS/PDR fusion algorithm specifically designed for smartwatches. This algorithm tracks the varied roll and pitch of the sensor caused by hand swings and integrates a CNN model to predict 1-D speed and perform ZUPT detection.

Projects

My projects span computer vision, robotics, and image processing, bridging theoretical research with practical applications across both academia and industry. I enjoy tackling challenging problems that require innovative solutions and have real-world impact.

	GPU IQ Enhancer Imaging Engineering team Logitech, 2025 video Developed a GPU-based IQ enhancer for Logitech's new Conference Cameras Products (Rally Board 65), featuring real-time face crop, image quality improvement, and noise reduction.
	CUDA-ISP Parallel Programming NYCU, 2024 code / report CUDA-ISP (Image Signal Processing) is a high-performance image processing pipeline designed for CUDA-enabled GPUs.
	Iris: Stereo camera localization in pre-build pointcloud maps POINT Lab & HDMap Center NCKU, 2023 code / video Stereo camera navigation in point cloud maps using ICP registration in challenging scenarios such as underground parking lots.
	Wrist-Worn IMU PDR Algorithm POINT Lab & AIROHA Technology NCKU, 2023 code A wrist-worn IMU PDR algorithm. Utilizing VQF to compute IMU attitude, divided into three stages: Step Detection, Step Length Estimation, and Heading Estimation, enabling navigation with Wrist-Worn IMU worn by pedestrians.
	Industrial Robot Controler Robotics Lab NCKU, 2022 code1 / code2 / video A ROS package for robot navigation and arm control. It is capable of recognizing and grasping objects, as well as navigating to any location on the map.
	Camera Navigation in Pre-build LiDAR Map POINT Lab NCKU, 2022 code / video A real-time camera navigation algorithm in a pre-built LiDAR map, utilizing NDT for 3D point cloud registration, effectively reducing 65% of accumulated position error.

Design and source code from Jon Barron's website.

Last updated Sep 2025.