Jinhyung (David) Park

I am a second-year PhD student at CMU's Robotics Institute, advised by Prof. Kris Kitani. I previously received my bachelor's degree in Computer Science at CMU in 2022, also working with Prof. Kris Kitani.

Additionally, I had the opportunity to conduct research at the MSC Lab in UC Berkeley for two summers, advised by Prof. Masayoshi Tomizuka and Dr. Wei Zhan. I also worked for a summer at Meta Zurich researching 3D panoptic reconstruction.

Email / CV / Scholar / Github

Research

I'm broadly interested in computer vision, joint 2D/3D understanding, and multi-modal learning. Much of my research focuses on bridging 2D and 3D representations for a cohesive understanding of the world.

	Flexible Depth Completion for Sparse and Varying Point Densities Jinhyung Park, Yu-Jhe Li, Kris Kitani CVPR, 2024 paper / bibtex Aligning predicted depth maps with observed depth points by propagating depth corrections improves depth completion for sparse and varying input point densities.
	Azimuth Super-Resolution for FMCW Radar in Autonomous Driving Yu-Jhe Li, Shawn Hunt, Jinhyung Park, Matthew O'Toole, Kris Kitani CVPR, 2023 paper / code / bibtex Super-resolution of radar using raw ADC signals effectively simulates additional receiver antennas and improves downstream detection performance.
	Time Will Tell: New Outlooks and A Baseline for Temporal Multi-View 3D Object Detection Jinhyung Park, Chenfeng Xu, Shijia Yang, Kurt Keutzer, Kris Kitani, Masayoshi Tomizuka, Wei Zhan ICLR, 2023 (Oral Presentation, Top 5% of accepted papers) paper / code / bibtex Combining long-term, low-resolution and short-term, high-resolution matching for temporal stereo yields efficient and performant camera-only 3D detectors.
	DetMatch: Two Teachers are Better Than One for Joint 2D and 3D Semi-Supervised Object Detection Jinhyung Park, Chenfeng Xu, Yiyang Zhou, Masayoshi Tomizuka, Wei Zhan ECCV, 2022 paper / code / bibtex Consistency between 2D and 3D pseudo-labels for joint 2D-3D semi-supervised learning stymies single-modality error propagation and improves performance.
	Modality-Agnostic Learning for Radar-Lidar Fusion in Vehicle Detection Yu-Jhe Li, Jinhyung Park, Matthew O'Toole, Kris Kitani CVPR, 2022 paper / bibtex Multi-modal fusion with prediction consistency between privileged teacher and noisy student alleivates collapse in difficult capture conditions and improves performance in ideal conditions.
	Multi-Modality Task Cascade for 3D Object Detection Jinhyung Park, Xinshuo Weng, Yunze Man, Kris Kitani BMVC, 2021 paper / bibtex Recursive, cascaded fusion of 2D and 3D representations at the task level improves both 2D segmentation and 3D detection quality.
	Crack Detection and Refinement Via Deep Reinforcement Learning Jinhyung Park, Yi-Chun Chen, Yu-Jhe Li, Kris Kitani ICIP, 2021 (Best Industry Impact Award) paper / bibtex Second-stage refinement of segmentation masks through an RL agent iteratively completes and cleans predictions.
	A Large-Scale Comprehensive Perception Dataset with High-Density Long-Range Point Clouds Xinshuo Weng, Dazhi Cheng, Yunze Man, Jinhyung Park, Matthew O'Toole, Kris Kitani arXiv, 2021 dataset page / paper / bibtex Large-scale synthetic driving dataset with comprehensive data distribution, sensor suite, and annotations.
	Protecting User Privacy: Obfuscating Discriminative Spatio-Temporal Footprints Jinhyung Park, Erik Seglem, Eric Lin, Andreas Züfle SIGSPATIAL LocalRec Workshop, 2017 paper / bibtex Consideration of entropy-based and adversarial obfuscation of user geolocation trajectories for online identity protection.
	Real-Time Bayesian Micro-Analysis for Metro Traffic Prediction Eric Lin, Jinhyung Park, Andreas Züfle SIGSPATIAL UrbanGIS Workshop, 2017 paper / bibtex Metro outflow prediction based on estimated distribution of origin-destination station pairs.

Design and source code from Jon Barron's website