Research teams from the laboratory of Prof. Kyungtae Kim in the Department of Electrical Engineering at POSTECH received two Best Paper Awards at the 2025 Summer Conference of the Korean Institute of Electromagnetic Engineering and Science (KIEES).

Seunghwan Kim (M.S. student) and Mingon Cho (Ph.D. candidate) were recognized for their research contributions, each receiving the Best Paper Award for separate studies presented at the conference.

Best Paper Award
Seunghwan Kim (M.S. student), Inhyuk Lee (Ph.D. candidate)
“A Modified RDA-Based Method for SFO Compensation in FMCW SAR”

This study proposes a technique for estimating and compensating for sampling frequency offset (SFO) that occurs in radar systems using different local oscillators. In particular, the research presents a signal model that incorporates SFO in an FMCW SAR environment, and develops a corresponding signal processing method and azimuth compression filter. Experimental results demonstrate that the proposed method can produce properly focused SAR images even in the presence of SFO, contributing to improved imaging quality in radar systems based on heterogeneous local oscillators.

Best Paper Award
Mingon Cho (Ph.D. candidate), Inhyuk Lee (Ph.D. candidate), Hyundong Kim (Ph.D. candidate), Otae Jang (Integrated M.S.–Ph.D. student)
“Point Scatterer Attention Module-Based Super-Resolution Network for SAR Imaging”

This study proposes a Point Scatterer Attention (PSA) module for SAR image super-resolution. The method encodes point scatterer (PS) locations estimated using the MUSIC algorithm into a binary mask and injects it into a U-Net architecture. The input consists of two channels—low-resolution SAR images and the PS mask—and a self-attention mechanism is designed using the averaged features of PS-activated regions as queries, guiding the reconstruction toward scatterer-centered structures. Experimental validation using MSTAR measured data shows that the PSA-based model outperforms baseline models across all evaluation metrics and reconstructs scatterer contours and key target structures more clearly. The results demonstrate the potential of the proposed method to improve imaging quality in applications such as ATR preprocessing and multistatic SAR systems.