Kwanghyun Koo

According to our database, Kwanghyun Koo authored at least 8 papers between 2023 and 2026.

Collaborative distances:

• Dijkstra number of five.
• Erdős number of five.

2026

SPEED: Structured kernel block pruning with filter groups for efficient and elastic SW-HW co-design in FPGA-based CNN accelerators.

Kwanghyun Koo, Sunwoong Kim, Hyun Kim

Neurocomputing, 2026

2025

TETRIS: On-Device Trainable Energy-Efficient FPGA Accelerator for Trustworthy and Real-Time Instance Segmentation.

Seungil Lee, Juntae Park, Kwanghyun Koo, Gilha Lee, Sangbeom Jeong, Junoh Park, Hyun Kim

Proceedings of the IEEE/ACM International Conference On Computer Aided Design, 2025

GASQ: Hardware-Friendly Gradient Distribution-Aware Split Quantization for Low-Bit CNN Training.

Sangbeom Jeong, Kwanghyun Koo, Hyun Kim

Proceedings of the International Conference on Electronics, Information, and Communication, 2025

2024

Vision transformer models for mobile/edge devices: a survey.

Seung Il Lee, Kwanghyun Koo, Jong Ho Lee, Gilha Lee, Sangbeom Jeong, Seongjun O, Hyun Kim

Multim. Syst., April, 2024

FB-SKP: Front-Back Structured Kernel Pruning via Group Convolution without Increasing Memory.

Kwanghyun Koo, Hyun Kim

Proceedings of the 21st International SoC Design Conference, 2024

Extreme Pruning Technique Based on Filter Deactivation Using Sparsity Training for Deep Convolutional Neural Networks.

Kwanghyun Koo, Hyun Kim

Proceedings of the International Conference on Electronics, Information, and Communication, 2024

2023

V-SKP: Vectorized Kernel-Based Structured Kernel Pruning for Accelerating Deep Convolutional Neural Networks.

Kwanghyun Koo, Hyun Kim

IEEE Access, 2023

PPT-KP: Pruning Point Training-based Kernel Pruning for Deep Convolutional Neural Networks.

Kwanghyun Koo, Hyun Kim

Proceedings of the 5th IEEE International Conference on Artificial Intelligence Circuits and Systems, 2023