Acoustic and vibration anomaly detection
Algorithms for surfacing early anomaly signatures in high-frequency acoustic and vibration data from motors, cranes, and pipes.
Research & patents
Signal-processing work that started in the lab now goes into industrial AI products built for deployment.
Algorithms for surfacing early anomaly signatures in high-frequency acoustic and vibration data from motors, cranes, and pipes.
Signal processing and interpretation for ultrasonic scanning of trees and structures, tracking internal change.
Speaker authentication and voice workflows designed to hold up in noisy industrial environments.
Model compression and optimization so AI runs in real time on FPGAs, MCUs, and industrial gateways. Training and validation run on in-house GPU and storage infrastructure, including 6 RTX 5090 Blackwell GPUs and 8 RTX 4090 Ada Lovelace GPUs.
2022
Seoul National University
2022
Korea Startup Promotion Agency (KISED)
2021
Seoul National University
2021
LG CNS
2020
Seoul National University
2020
Korea Forest Research Institute
2019
Seoul National University