Hyo Jin Kim

Research Scientist · Reality Labs Research, Meta

Google Scholar LinkedIn

About

I am a research scientist at Reality Labs Research, Meta. My current work focuses on egocentric contextual AI for always-on wearable devices such as smart glasses — long-horizon video understanding and multimodal models grounded in a user's everyday visual stream. Earlier, my research centered on visual localization, learned image descriptors, scene understanding, and the privacy and security of visual data in AR / VR systems.

I received my Ph.D. in Computer Science from the University of North Carolina at Chapel Hill, advised by Prof. Jan-Michael Frahm. Prior to that, I obtained M.S. degree in EECS at Seoul National University under the supervision of Prof. Kyoung Mu Lee, and B.S. in Electrical Engineering at KAIST.

Publications

Recent

• Agentic Very Long Video Understanding
  Aniket Rege, Arka Sadhu, Yuliang Li, Kejie Li, Ramya Korlakai Vinayak, Yuning Chai, Yong Jae Lee, Hyo Jin Kim (corresponding author)
  ACL 2026 (Main) · Paper · Project page · Code
• SuperMemory-VQA: An Egocentric Visual Question-Answering Benchmark for Long-Horizon Memory
  Samiul Alam, Shakhrul Iman Siam, Michael J. Proulx, James Fort, Richard Newcombe, Hyo Jin Kim, Mi Zhang
  arXiv preprint · Paper · Project page · Code
• Reading Recognition in the Wild
  Charig Yang, Samiul Alam, Shakhrul Iman Siam, Michael J. Proulx, Lambert Mathias, Kiran Somasundaram, Luis Pesqueira, James Fort, Sheroze Sheriffdeen, Omkar Parkhi, Carl Ren, Mi Zhang, Yuning Chai, Richard Newcombe, Hyo Jin Kim (corresponding author)
  NeurIPS 2025 · Paper · Dataset · Data Explorer · Code
• EgoLM: Multi-Modal Language Model of Egocentric Motions
  Fangzhou Hong, Vladimir Guzov, Hyo Jin Kim, Yuting Ye, Richard Newcombe, Ziwei Liu, Lingni Ma
  CVPR 2025 · Oral · Paper · Project page
• Nymeria: A Massive Collection of Multimodal Egocentric Daily Motion in the Wild
  Lingni Ma, Yuting Ye, Fangzhou Hong, Vladimir Guzov, Yifeng Jiang, Rowan Postyeni, Luis Pesqueira, Alexander Gamino, Vijay Baiyya, Hyo Jin Kim, Kevin Bailey, David S. Fosas, C. Karen Liu, Ziwei Liu, Jakob Engel, Renzo De Nardi, Richard Newcombe
  ECCV 2024 · Paper · Dataset

Selected earlier work

• NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning
  Tony Ng, Hyo Jin Kim (corresponding author), Vincent T. Lee, Daniel DeTone, Tsun-Yi Yang, Tianwei Shen, Eddy Ilg, Vassileios Balntas, Krystian Mikolajczyk, Chris Sweeney
  CVPR 2022 · Paper · Code
• Analysis and Mitigations of Reverse Engineering Attacks on Local Feature Descriptors
  Deeksha Dangwal, Vincent T. Lee, Hyo Jin Kim, Tianwei Shen, Meghan Cowan, Rajvi Shah, Caroline Trippel, Brandon Reagen, Timothy Sherwood, Vasileios Balntas, Armin Alaghi, Eddy Ilg
  BMVC 2021 · Paper
• Domain Adaptation of Learned Features for Visual Localization
  Sungyong Baik, Hyo Jin Kim (corresponding author), Tianwei Shen, Eddy Ilg, Kyoung Mu Lee, Chris Sweeney
  BMVC 2020 · Paper
• Hierarchy of Alternating Specialists for Scene Recognition
  Hyo Jin Kim, Jan-Michael Frahm
  ECCV 2018 · Paper
• Learned Contextual Feature Reweighting for Image Geo-Localization
  Hyo Jin Kim, Enrique Dunn, Jan-Michael Frahm
  CVPR 2017 · Spotlight · Paper · Supplementary · Poster · Code · Video
• Predicting Good Features for Image Geo-Localization Using Per-Bundle VLAD
  Hyo Jin Kim, Enrique Dunn, Jan-Michael Frahm
  ICCV 2015 · Paper · Video

A full list is available on my Google Scholar profile.

Talks

• Privacy Challenges for Machine Perception in AR
  Project Aria Tutorial, CVPR 2022 · Video
• Symmetry-Growing for Skewed Rotational Symmetry Detection
  Hyo Jin Kim, Minsu Cho, Kyoung Mu Lee
  Symmetry Detection in Real-World Images – A Competition, CVPR 2011 · Winner, Rotation Symmetry Detection

Service

Reviewer — TPAMI, CVPR, ICCV, ECCV, AAAI, ICLR, BMVC, ACCV, 3DV, WACV

Media

• Women in Computer Vision: Hyo Jin Kim
  Computer Vision News, November 2019, pp. 30–33 · Link