Sung-Un Park
박성언 (朴成彦)
Staff Engineer · Samsung Electronics
Vertical AI1 Team, AX/PI Center
Industrial AI · Computer Vision · On-Device AI
Suwon, Republic of Korea
park@stoz.kr · LinkedIn · Google Scholar · Naver
Embedded Files
About
I am a Staff Engineer in the Vertical AI1 Team at Samsung Electronics’ AX/PI Center. My current work focuses on virtual metrology, using process data to predict metrology values and support the detection of abnormal process conditions. Previously, I worked on SEM-based depth estimation and biometric systems for Galaxy devices, including fingerprint anti-spoofing, speaker recognition, and face recognition. My research background also includes human activity recognition from depth images and machine learning for biomedical applications. I am pursuing a Ph.D. in Electrical Engineering at KAIST.
Research and Industry Experience
Samsung Electronics
Suwon, Republic of Korea | Feb 2017–Present
Current role: Staff Engineer, Vertical AI1 Team, AX/PI Center
2026–Present
Previously a Staff Researcher in the Process/Equipment AI Lab, AI Development, AI Center. Earlier affiliation: AI Research Center, Samsung Advanced Institute of Technology (SAIT).
Virtual Metrology
May 2024–Present
Develop machine learning models to predict metrology values from semiconductor process data for process monitoring.
3D Depth Metrology for Semiconductors
Jan 2022–Jun 2024
Developed deep learning algorithms to estimate the depth of semiconductor structures from scanning electron microscope (SEM) images for 3D metrology.
Related public resource: Semiconductor SEM competition.
Fingerprint Anti-Spoofing for the Galaxy S Series
Sep 2019–Sep 2022
Designed the architecture of an on-device fingerprint anti-spoofing system deployed on Galaxy devices.
Developed spoof detection algorithms for diverse user conditions.
Related granted patent: US11,935,321 B2.
Bixby Wake-Up and Unlock: Speaker Recognition for Galaxy Devices
Jul 2018–Aug 2019
Designed and implemented the core software architecture of the Bixby speaker recognition system.
Developed algorithms to improve speaker recognition across noise conditions and update user enrollment data for more robust recognition.
Related granted patents: US11,289,098 B2 · US11,545,154 B2.
Product reference: Bixby on the Galaxy S10.
Intelligent Scan for the Galaxy S9 and Note9
Aug 2017–Jun 2018
Contributed to the software architecture of Intelligent Scan, which combines iris scanning and face recognition.
Developed multimodal face liveness detection algorithms for anti-spoofing.
Developed landmark alignment for weakly synchronized images with different fields of view, resolutions, and spectral characteristics, reducing detection latency.
Streamlined processing to reduce face recognition time and implemented image filtering and preprocessing for low-light conditions.
Investigated hard sample mining strategies to improve deep learning robustness.
Related granted patents: US11,023,757 B2 · US10,853,678 B2 · US11,625,473 B2 · US11,682,232 B2 · US11,449,971 B2 · EP3,525,165 B1.
Product references: Galaxy S9 security features · Galaxy S9 feature overview.
Face Recognition for the Galaxy S8 and Note8
Mar 2017–Jul 2017
Contributed to face recognition system development and integrated recognition software into the mobile system for product release.
Kyung Hee University — Bio-Imaging Lab
Research Experience
Yongin, Republic of Korea | May 2013–Jan 2017
Human Activity Recognition from Depth Images — Thesis Research
Developed a system to recognize activities such as walking, jumping, and raising arms from depth image sequences.
Used joint-angle features and recurrent neural networks implemented in TensorFlow.
TFT-LCD Panel Defect Detection and Classification — Samsung Display
Jul–Dec 2016
Developed a convolutional neural network (CNN) system for automated detection and classification of manufacturing defects in TFT-LCD panels.
Implemented real-time detection and classification.
Human Image Retrieval — Mappers Corporation
Nov 2015–Mar 2016
Developed a CNN-based system to classify and retrieve images containing people.
Automatic Sleep Stage Classification — Samsung Electronics, Home Appliances Division, and Kyung Hee University Hospital at Gangdong
2015
Developed an algorithm for sleep stage classification using six-channel EEG, EMG, and EOG data.
Achieved 87% accuracy against manual scoring by clinical experts.
Samsung Software Membership — Samsung Electronics
Software Engineer
Suwon, Republic of Korea | Jan 2013–Mar 2015
Developed applications providing technical barriers to trade (TBT) information and harmonized system (HS) code lookup; received an award for excellent development in the 2015 Technical Standard Idea Software Contest.
Developed a collaborative online sticky-note application with synchronized viewing and editing for brainstorming (2014; funded by Samsung Electronics, Visual Display Business). Demo
Developed an automatic response system (ARS) speech-to-text application to display spoken prompts; received an Excellent Project award at the Samsung Software Membership Technology Fair 2014. Demo
Developed an Android MAME emulator supporting multiplayer gameplay over Wi-Fi Direct (2013).
Evaluated the usability of the Tizen 2.2.1 SDK and performed Tizen compliance testing (2013; funded by Samsung Electronics, IT & Mobile Business).
Developed “Easy Travel,” an application offering automated itinerary planning and social features for Nailo train travelers (2013; supported by SK planet’s Open Platform group).
Education
Korea Advanced Institute of Science and Technology (KAIST)
Ph.D. Student in Electrical Engineering
Daejeon, Republic of Korea | Aug 2021–Present
Coursework requirements completed; degree in progress.
Advisor: Prof. Changick Kim
Kyung Hee University
M.S. in Biomedical Engineering
Yongin, Republic of Korea | Mar 2015–Feb 2017
Thesis: “A Depth Camera-based Human Activity Recognition via Deep Learning Recurrent Neural Network”
Advisor: Prof. Tae-Seong Kim
B.S. in Biomedical Engineering and B.S. in Computer Engineering
Yongin, Republic of Korea | Mar 2009–Feb 2015
Entered as an undeclared student.
Student Council President, College of Liberal Arts, Kyung Hee University (2009).
Completed mandatory military service in the Republic of Korea Army during this period (Oct 2010–Jul 2012); honorably discharged as a Sergeant.
Samsung Software Membership, 23rd cohort (2013–2015).
Publications
International Journals and Conferences
PRIME: Deep Imbalanced Regression with Proxies.
J. Lim, S. Lee, D. Um, S. U. Park, and J. Shin.
International Conference on Machine Learning (ICML), 2025. Paper
Robust Speaker Verification with Joint Self-Supervised and Supervised Learning.
K. Wang, X. Zhang, M. Zhang, Y. Li, J. Lee, K. Cho, and S. U. Park.
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022. DOI
End-to-End Transformer-Based Open-Vocabulary Keyword Spotting with Location-Guided Local Attention.
B. Wei, M. Yang, T. Zhang, X. Tang, X. Huang, K. Kim, J. Lee, K. Cho, and S. U. Park.
Interspeech, pp. 361–365, 2021.
An Automatic Computer-Aided Diagnosis System for Breast Cancer in Digital Mammograms via Deep Belief Network.
M. A. Al-antari, M. A. Al-masni, S. U. Park, J. H. Park, M. K. Metwally, Y. M. Kadah, S. M. Han, and T.-S. Kim.
Journal of Medical and Biological Engineering, vol. 38, pp. 443–456, 2018. DOI
A Depth Camera-based Human Activity Recognition via Deep Learning Recurrent Neural Network for Health and Social Care Services.
S. U. Park, J. H. Park, M. A. Al-masni, M. A. Al-antari, Md. Z. Uddin, and T.-S. Kim.
Conference on Health and Social Care Information Systems and Technologies (HCist), 2016. Paper
Automatic Computer-Aided Diagnosis of Breast Cancer in Digital Mammograms via Deep Belief Network.
M. A. Al-masni, S. U. Park, J. H. Park, Y. M. Kadah, S. M. Han, T.-S. Kim, and M. A. Al-antari.
Global Conference on Engineering and Applied Science, pp. 1306–1314, 2016.
A Single Depth Sensor Based Human Activity Recognition via Convolutional Neural Network.
J. H. Park, S. U. Park, M. Z. Uddin, M. A. Al-antari, M. A. Al-masni, and T.-S. Kim.
6th International Conference on the Development of Biomedical Engineering, pp. 329–332, 2016.
Accurate 3D Human Pose Recognition via Fusion of Depth and Motion Sensors.
S. B. Nam, S. U. Park, J. H. Park, M. Z. Uddin, and T.-S. Kim.
International Journal of Future Computer and Communication, vol. 4, p. 336, 2015.
A Single Depth Sensor Based Human Activity Recognition via Deep Belief Network.
S. B. Nam, S. U. Park, J. H. Park, and T.-S. Kim.
4th World Conference on Applied Sciences, Engineering and Technology, pp. 15–19, 2015.
Unsupervised 3D human pose recognition from a single depth human silhouette using a geodesic map and kinematic body model.
H. J. Jeon, S. B. Nam, S. U. Park, J. H. Park, C. Yoo, J. T. Kim, and T. Kim.
9th International Conference on Ubiquitous Information Management and Communication, p. 70, 2015.
Korean Society of Medical & Biological Engineering Proceedings
A novel computer-aided diagnosis system for breast cancer in digital mammograms via deep belief network.
M. A. Al-antari, M. A. Al-masni, S. U. Park, J. H. Park, Y. M. Kadah, S. M. Han, and T.-S. Kim.
May 2016, 51, pp. 397–398.
Extraction of Retinal Blood Vessels Utilizing 2D Matched Filter.
M. A. Al-antari, M. A. Al-masni, H. M. Bu-Omer, A. M. Anam, S. U. Park, J. H. Park, N. H. Solouma, S. M. Han, and T.-S. Kim.
May 2016, 51, pp. 303–304.
Human Activity Recognition with a Depth Imaging Sensor via Convolutional Neural Network.
J. H. Park, S. U. Park, J. M. Park, M. A. Al-antari, M. A. Al-masni, and T.-S. Kim.
May 2016, 51, pp. 46–47.
Improved 3D human skeletal pose reconstruction using the probability distribution of random forests in body parts recognition and joint proposals estimation.
S. B. Nam, S. U. Park, H. S. Han, H. J. Jeon, S. M. Kim, and T.-S. Kim.
Nov 2013, 48, pp. 724–725.
Patents and Published Applications
Country codes indicate associated jurisdictions; grant status varies by jurisdiction.
Granted U.S. Patents
Method and apparatus with biometric information spoof detection
Sung-Jae Cho, Kyuhong Kim, Sungun Park, Jiwhan Kim, and JaeJoon HAN
US12321445B2 · 2025
Jurisdictions: US; KR; CN; EP
Method and apparatus with image fusion
Minsu Ko, SeungJu HAN, JaeJoon HAN, Jihye Kim, Sungun Park, and Chang Kyu Choi
US12315115B2 · 2025
Jurisdictions: US; KR; CN; EP
Method and apparatus with anti-spoofing
Sungun Park and Kyuhong Kim
US12307807B2 · 2025
Jurisdictions: US; KR; CN; EP; JP
Method and apparatus with image preprocessing
Jaeyun Lee, Kyuhong Kim, Sungun Park, and JaeJoon HAN
US12299851B2 · 2025
Jurisdictions: US; KR
Method and apparatus with input data classification
Jaeyun Lee, Kyuhong Kim, Jiwhan Kim, Sungun Park, Kiho Cho, and JaeJoon HAN
US12229184B2 · 2025
Jurisdictions: US; KR
Anti-spoofing method and apparatus
Sungun Park, Kyuhong Kim, and JaeJoon HAN
US12147516B2 · 2024
Jurisdictions: US; KR; CN; EP
Method and apparatus with biometric spoofing consideration
Jiwhan Kim, Kyuhong Kim, Sungun Park, and JaeJoon HAN
US12087096B2 · 2024
Jurisdictions: US; KR; CN; EP
Device and method with image matching
Sungun Park, Jihye Kim, JaeJoon HAN, Minsu Ko, SeungJu HAN, Jinwoo SON, and Changyong Son
US12062251B2 · 2024
Jurisdictions: US; KR; CN; EP
Method and apparatus with liveness verification
Sungun Park, Hyeongwook YANG, Tushar Balasaheb Sandhan, Jiaqian YU, Jingtao XU, Youngjun KWAK, Juwoan YOO, and Jae Joon Han
US12014571B2 · 2024
Jurisdictions: US; KR
Method and apparatus that detects spoofing of biometric information
Sung-Jae Cho, Kyuhong Kim, Sungun Park, Geuntae BAE, and JaeJoon HAN
US11915525B2 · 2024
Jurisdictions: US; KR; CN; EP
Method and apparatus with anti-spoofing
Sungun Park and Kyuhong Kim
US11935321B2 · 2024
Jurisdictions: US; KR; CN; EP; JP
Device and method with image matching
Sungun Park, Jihye Kim, JaeJoon HAN, Minsu Ko, SeungJu HAN, Jinwoo SON, and Changyong Son
US11682232B2 · 2023
Jurisdictions: US; KR; CN; EP
Method and apparatus with selective combined authentication
Jihye Kim, SeungJu HAN, JaeJoon HAN, Minsu Ko, Sungun Park, and Chang Kyu Choi
US11625473B2 · 2023
Jurisdictions: US; KR; CN; EP
Method and device with data recognition
Sung-un Park and Kyuhong Kim
US11574641B2 · 2023
Jurisdictions: US; KR
Apparatus with fake fingerprint detection and method with updating of model to prevent fake fingerprint
Sung-Jae Cho, Kyuhong Kim, and Sungun Park
US11574499B2 · 2023
Jurisdictions: US; KR
Method and apparatus with registration for speaker recognition
Sung-un Park and Kyuhong Kim
US11545154B2 · 2023
Jurisdictions: US; KR
Method and apparatus with image fusion
Minsu Ko, SeungJu HAN, JaeJoon HAN, Jihye Kim, Sungun Park, and Chang Kyu Choi
US11449971B2 · 2022
Jurisdictions: US; KR; CN; EP
Object recognition method and apparatus
SeungJu HAN, Sungun Park, JaeJoon HAN, Jinwoo SON, Changyong Son, Minsu Ko, Jihye Kim, and Chang Kyu Choi
US11423702B2 · 2022
Jurisdictions: US; KR; CN; EP; JP
Method and apparatus with fake fingerprint detection
Jiwhan Kim, Sungun Park, Kyuhong Kim, and Geuntae BAE
US11335117B2 · 2022
Jurisdictions: US; KR
Method and apparatus with speaker recognition registration
Sung-un Park and Kyuhong Kim
US11289098B2 · 2022
Jurisdictions: US; KR; CN; EP
Method and apparatus with liveness verification
Sungun Park, Hyeongwook YANG, Tushar Balasaheb Sandhan, Jiaqian YU, Jingtao XU, Youngjun KWAK, Juwoan YOO, and Jae Joon Han
US11023757B2 · 2021
Jurisdictions: US; KR
Object recognition method and apparatus
SeungJu HAN, Sungun Park, JaeJoon HAN, Jinwoo SON, Changyong Son, Minsu Ko, Jihye Kim, and Chang Kyu Choi
US10853678B2 · 2020
Jurisdictions: US; KR; CN; EP; JP
Published Patent Applications
Method and apparatus with AI model training using domain similarity
Junhyun NAM, Minsu KO, Sungun Park, Sungjoo Suh, and Minsu Ahn
Application EP 25150247.2 · EP4583005A1 · 2025
Jurisdictions: EP; KR; CN
Method and apparatus with scanning electron microscope image correction
Sungun Park, Minsu Ko, Junhyun NAM, In Seok Park, Sung Yoon Ryu, Sungjoo Suh, and Minsu Ahn
Application US 18/982,554 · US20250218041A1 · 2025
Jurisdictions: US; KR
Method and device with neural network
Jaeyun Lee, Kiho Cho, Jiwhan Kim, Sungun Park, and JaeJoon HAN
Application US 17/982,618 · US20230252283A1 · 2023
Jurisdictions: US; KR
Honors and Awards
SAIT Vision Awards — Best Challenge, Samsung Advanced Institute of Technology, Aug 13, 2024; first half of 2024.
Samsung Best Paper Award — Silver Award, Samsung Electronics, Oct 20, 2021. “Fast and Robust Fingerprint Anti-Spoofing for Mobile Devices via Joint Contrastive Learning.”
Samsung Best Paper Award — Bronze Award, Samsung Electronics, Oct 2019. “Bayesian Speaker Verification using Orthogonal Deep Representation.”
Future Creator Award, Device Solutions Division, Samsung Electronics, Apr 25, 2019; first quarter of 2019.
Boundless Search for Breakthroughs Award (Innovation), Samsung Advanced Institute of Technology, Jan 3, 2018; second half of 2017.
Excellent Development Award, Technical Standard Idea Software Contest, Korean Agency for Technology and Standards, Apr 2015.
Excellent Project, Samsung Software Membership Technology Fair, Samsung Electronics, Sep 2014. Selected as a regional representative team.
Technical Skills
Areas: Industrial AI, computer vision, machine learning, on-device AI, biometric security
Programming: C++, Python, C
Additional languages: Java, JavaScript, PHP, SQL, MATLAB
Frameworks and libraries: PyTorch; TensorFlow used in human activity recognition research; MFC
Web technologies: HTML
Additional Activities
Music: If you’re just like me.
In memoria aeterna erit iustus. ab auditione mala non timebit.
Report abuse