Pei-Yin Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Pei-Yin Chen has a total of 103 co-authors. Here are the top co-authors by number of shared publications:
- Kuan-Yu Huang
Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan 70101, Taiwan
7 shared publications - Wei-Ting Chen
MediaTek Inc., Hsinchu, Taiwan
6 shared publications Chih-Yuan Lien
Department of Electronic Engineering, National Kaohsiung University of Sciences and Technology, Kaohsiung, Taiwan
4 shared publications- Shih-Hsiang Lin
Digital IC Design Laboratory, Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan 70101, Taiwan
4 shared publications - Yao-Tsung Kuo
Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan, Taiwan
4 shared publications - Wen-Long Chin
Department of Engineering Science, National Cheng Kung University, Tainan 701, Taiwan
3 shared publications
Latest Publications
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 6982-6996
Bridging Training–Deployment Gap in Intrusion Detection With Source-Free Domain Adaptation
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 6982-6996
Bridging Training–Deployment Gap in Intrusion Detection With Source-Free Domain Adaptation
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 103185-103194
High-Efficiency Double SHA-256 Architecture via Expansion–Compression Co-Design
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 103185-103194
Open AccessHigh-Efficiency Double SHA-256 Architecture via Expansion–Compression Co-Design
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 100351-100363
A Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 100351-100363
Open AccessA Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains
- 18 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (2) , 746-750
CRC-Based Error Detection Mechanism for Multiplicative Inverse in Redundant Basis S -Box
advanced encryption standard (AES) is the encryption system that ensures the data confidentiality of wireless communication between the satellite and the ground station. For space missions in harsh radiation environments, satellite equipment must be equipped with a robust and highly reliable encryption system to ensure secure communication and operations. In this brief, a CRC-based error detection mechanism for multiplicative inverse in redundant-basis AES S-Box is proposed. The architecture is synthesized by Synopsys Design Compiler with Nangate 45-nm Open Cell Library. The implementation results show that the proposed mechanisms achieve 99.993% error coverage, demonstrating their strong suitability.
- 18 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (2) , 746-750
CRC-Based Error Detection Mechanism for Multiplicative Inverse in Redundant Basis S -Box
advanced encryption standard (AES) is the encryption system that ensures the data confidentiality of wireless communication between the satellite and the ground station. For space missions in harsh radiation environments, satellite equipment must be equipped with a robust and highly reliable encryption system to ensure secure communication and operations. In this brief, a CRC-based error detection mechanism for multiplicative inverse in redundant-basis AES S-Box is proposed. The architecture is synthesized by Synopsys Design Compiler with Nangate 45-nm Open Cell Library. The implementation results show that the proposed mechanisms achieve 99.993% error coverage, demonstrating their strong suitability.
- Open Access
QP-Adaptive Dual-Path Residual Integrated Frequency Transformer for Data-Driven In-Loop Filter in VVC
As AI-enabled embedded systems such as smart TVs and edge devices demand efficient video processing, Versatile Video Coding (VVC/H.266) becomes essential for bandwidth-constrained Multimedia Internet of Things (M-IoT) applications. However, its block-based coding often introduces compression artifacts. While CNN-based methods effectively reduce these artifacts, maintaining robust performance across varying quantization parameters (QPs) remains challenging. Recent QP-adaptive designs like QA-Filter show promise but are still limited. This paper proposes DRIFT, a QP-adaptive in-loop filtering network for VVC. DRIFT combines a lightweight frequency fusion CNN (LFFCNN) for local enhancement and a Swin Transformer-based global skip connection for capturing long-range dependencies. LFFCNN leverages octave convolution and introduces a novel residual block (FFRB) that integrates multiscale extraction, QP adaptivity, frequency fusion, and spatial-channel attention. A QP estimator (QPE) is further introduced to mitigate double enhancement in inter-coded frames. Experimental results demonstrate that DRIFT achieves BD rate reductions of 6.56% (intra) and 4.83% (inter), with an up to 10.90% gain on the BasketballDrill sequence. Additionally, LFFCNN reduces the model size by 32% while slightly improving the coding performance over QA-Filter.
Open AccessQP-Adaptive Dual-Path Residual Integrated Frequency Transformer for Data-Driven In-Loop Filter in VVC
As AI-enabled embedded systems such as smart TVs and edge devices demand efficient video processing, Versatile Video Coding (VVC/H.266) becomes essential for bandwidth-constrained Multimedia Internet of Things (M-IoT) applications. However, its block-based coding often introduces compression artifacts. While CNN-based methods effectively reduce these artifacts, maintaining robust performance across varying quantization parameters (QPs) remains challenging. Recent QP-adaptive designs like QA-Filter show promise but are still limited. This paper proposes DRIFT, a QP-adaptive in-loop filtering network for VVC. DRIFT combines a lightweight frequency fusion CNN (LFFCNN) for local enhancement and a Swin Transformer-based global skip connection for capturing long-range dependencies. LFFCNN leverages octave convolution and introduces a novel residual block (FFRB) that integrates multiscale extraction, QP adaptivity, frequency fusion, and spatial-channel attention. A QP estimator (QPE) is further introduced to mitigate double enhancement in inter-coded frames. Experimental results demonstrate that DRIFT achieves BD rate reductions of 6.56% (intra) and 4.83% (inter), with an up to 10.90% gain on the BasketballDrill sequence. Additionally, LFFCNN reduces the model size by 32% while slightly improving the coding performance over QA-Filter.
- 25 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (3) , 4869-4881
MTBWD: Multidirectional Two-Step Bisection Weighting Demosaicking for Image Sensors
This paper proposes a multidirectional two-step bisection weighting demosaicking (MTBWD) method for the image sensors. For image demosaicking, complicated operations are typically employed to calculate more precise weights and reconstructed pixel values. Considering hardware implementation, we tend to avoid using complicated operations, such as multiplication. Therefore, bisection weighting is proposed, implying that all interpolated weights are designed to be the fraction with a value of denominator that is a power of 2. However, if we only use the bisection weighting, it is challenging to adopt the multidirectional information. Therefore, we proposed two-step bisection weighting. In the first step, we compute the global directional information through the first bisection weighting of local directional information, consisting of variations and color difference values of cardinal and oblique directions. Then, the final color difference value is computed through the second-time bisection weighting of global directional information to interpolate the missing pixel. To reduce hardware cost, under the limitation of avoiding multiplication, MTBWD could still utilize multidirectional information to achieve competitive image quality. The very large-scale integration (VLSI) architecture of MTBWD is implemented and synthesized using TSMC 40-nm technology. Experimentally, compared with the related low-cost methods, the proposed MTBWD achieves the best average quality for diverse images, including images in Kodak24, IMAX, and DIV2K datasets and some photos captured in a specific scene. The hardware reports confirm that the MTBWD is the most cost-effective in the current low-cost demosaicking studies with similar hardware costs.
- 25 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (3) , 4869-4881
MTBWD: Multidirectional Two-Step Bisection Weighting Demosaicking for Image Sensors
This paper proposes a multidirectional two-step bisection weighting demosaicking (MTBWD) method for the image sensors. For image demosaicking, complicated operations are typically employed to calculate more precise weights and reconstructed pixel values. Considering hardware implementation, we tend to avoid using complicated operations, such as multiplication. Therefore, bisection weighting is proposed, implying that all interpolated weights are designed to be the fraction with a value of denominator that is a power of 2. However, if we only use the bisection weighting, it is challenging to adopt the multidirectional information. Therefore, we proposed two-step bisection weighting. In the first step, we compute the global directional information through the first bisection weighting of local directional information, consisting of variations and color difference values of cardinal and oblique directions. Then, the final color difference value is computed through the second-time bisection weighting of global directional information to interpolate the missing pixel. To reduce hardware cost, under the limitation of avoiding multiplication, MTBWD could still utilize multidirectional information to achieve competitive image quality. The very large-scale integration (VLSI) architecture of MTBWD is implemented and synthesized using TSMC 40-nm technology. Experimentally, compared with the related low-cost methods, the proposed MTBWD achieves the best average quality for diverse images, including images in Kodak24, IMAX, and DIV2K datasets and some photos captured in a specific scene. The hardware reports confirm that the MTBWD is the most cost-effective in the current low-cost demosaicking studies with similar hardware costs.
- Open Access
- 1 August 2024
- Published by Institution of Engineering and Technology (IET) in IET Smart Grid
- Vol. 7 (6) , 872-890
Application of a three‐phase unified power quality conditioner in a microgrid
In power distribution systems, common issues such as voltage fluctuations, voltage instability, current harmonics, and power imbalances often arise, negatively impacting the stability and power quality (PQ) of the power system. As an advanced power electronics device, the unified power quality conditioner (UPQC) can effectively address these problems. Therefore, the integration of UPQC into power distribution lines can enhance the stability and reliability of the power system, providing higher‐quality power supply to meet the needs of both users and suppliers for PQ. The authors utilise the TI TMS320F28377D chip as the control core to construct a simple yet highly efficient UPQC system. Through simulation, the correctness of the voltage and current compensation algorithms has been verified, and the overall system feasibility has been validated using a 100‐kVA series‐shunt inverter. The experimental results show that the voltage compensation range of the UPQC is 380V ± 10%. During sudden current drops, the imbalance rate can be reduced from 85.5% to 1.32% after controlled compensation. During sudden current rises, the imbalance rate can be reduced from 13.1% to 1.56% after controlled compensation. These findings demonstrate the effectiveness of the UPQC system in improving PQ and system stability in power distribution networks.
- 1 August 2024
- Published by Institution of Engineering and Technology (IET) in IET Smart Grid
- Vol. 7 (6) , 872-890
Open AccessApplication of a three‐phase unified power quality conditioner in a microgrid
In power distribution systems, common issues such as voltage fluctuations, voltage instability, current harmonics, and power imbalances often arise, negatively impacting the stability and power quality (PQ) of the power system. As an advanced power electronics device, the unified power quality conditioner (UPQC) can effectively address these problems. Therefore, the integration of UPQC into power distribution lines can enhance the stability and reliability of the power system, providing higher‐quality power supply to meet the needs of both users and suppliers for PQ. The authors utilise the TI TMS320F28377D chip as the control core to construct a simple yet highly efficient UPQC system. Through simulation, the correctness of the voltage and current compensation algorithms has been verified, and the overall system feasibility has been validated using a 100‐kVA series‐shunt inverter. The experimental results show that the voltage compensation range of the UPQC is 380V ± 10%. During sudden current drops, the imbalance rate can be reduced from 85.5% to 1.32% after controlled compensation. During sudden current rises, the imbalance rate can be reduced from 13.1% to 1.56% after controlled compensation. These findings demonstrate the effectiveness of the UPQC system in improving PQ and system stability in power distribution networks.
- Open Access
- 24 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 88596-88607
Battery Management System for Electric Garbage Compactor Trucks
When garbage trucks perform garbage collection and compression operations, it is common to keep the engine idling and even increase the engine speed during garbage compression, which can lead to noise, air pollution, increased fuel consumption, and carbon emissions. Adopting an electric compression system can effectively reduce these issues. The battery pack used in electric garbage trucks is the core energy source of the vehicle, making proper battery management system crucial for the overall safety and performance of the vehicle. This study aims to utilize NUVOTON’s Cortex-M4 chip to develop a battery management system specifically designed for electric garbage trucks. By real-time online estimation of the battery state, optimal performance of the battery pack can be achieved. Battery health is assessed based on capacity cycle counting for parameter weighting evaluation of battery voltage drop. By comparing the capacities of battery modules to track and calibrate the open-circuit voltage, the capacity error is primarily estimated using a combination of Coulomb counting method and open-circuit voltage method to assess the battery’s state of charge and evaluate its lifespan. The proposed method is validated by integrating the battery state estimation technique into the microcontroller of the battery management system, and compared with the conventional Coulomb counting method. The real-time online battery estimation method adjusts the initial value check of SoC by tracking the variation of battery module capacity and adjusting the OCV lookup table, thereby enhancing the accuracy of SoC estimation and reducing errors. This method can be effectively applied to electric garbage compressors to improve battery utilization efficiency and maximize battery lifespan.
- 24 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 88596-88607
Open AccessBattery Management System for Electric Garbage Compactor Trucks
When garbage trucks perform garbage collection and compression operations, it is common to keep the engine idling and even increase the engine speed during garbage compression, which can lead to noise, air pollution, increased fuel consumption, and carbon emissions. Adopting an electric compression system can effectively reduce these issues. The battery pack used in electric garbage trucks is the core energy source of the vehicle, making proper battery management system crucial for the overall safety and performance of the vehicle. This study aims to utilize NUVOTON’s Cortex-M4 chip to develop a battery management system specifically designed for electric garbage trucks. By real-time online estimation of the battery state, optimal performance of the battery pack can be achieved. Battery health is assessed based on capacity cycle counting for parameter weighting evaluation of battery voltage drop. By comparing the capacities of battery modules to track and calibrate the open-circuit voltage, the capacity error is primarily estimated using a combination of Coulomb counting method and open-circuit voltage method to assess the battery’s state of charge and evaluate its lifespan. The proposed method is validated by integrating the battery state estimation technique into the microcontroller of the battery management system, and compared with the conventional Coulomb counting method. The real-time online battery estimation method adjusts the initial value check of SoC by tracking the variation of battery module capacity and adjusting the OCV lookup table, thereby enhancing the accuracy of SoC estimation and reducing errors. This method can be effectively applied to electric garbage compressors to improve battery utilization efficiency and maximize battery lifespan.
- 28 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (2) , 717-730
A Low-Cost Pipelined Architecture Based on a Hybrid Sorting Algorithm
In this paper, a low-cost pipelined architecture based on a hybrid sorting algorithm is proposed. The proposed architecture is constructed with a bitonic sorter and several cascaded bidirectional insertion sorting units. The bidirectional insertion sorting unit uses the segmented sorted subsequence generated by the bitonic sorter as input, and records the maximum and minimum values of the subsequence. After all segmented subsequences are processed through the cascaded bidirectional insertion sorting units, a sorted sequence is obtained. The proposed architecture is implemented using the Verilog hardware description language (HDL) and synthesized using the Synopsys Design Compiler with a TSMC 90-nm cell library. The experimental results indicate that the proposed architecture can not only shorten sorting cycles but also reduce hardware area costs. Moreover, sorting cycles can be further shortened by increasing the parallelism of the proposed architecture. Under the configuration that 2048 32-bit data to be sorted and 16 data have to be processed simultaneously, the proposed architecture can improve the throughput-to-gate-count ratio by 16%, and throughput-to-power-consumption-ratio by 25% compared to the existing sorting design. The proposed architecture makes the most efficient use of hardware resources.
- 28 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (2) , 717-730
A Low-Cost Pipelined Architecture Based on a Hybrid Sorting Algorithm
In this paper, a low-cost pipelined architecture based on a hybrid sorting algorithm is proposed. The proposed architecture is constructed with a bitonic sorter and several cascaded bidirectional insertion sorting units. The bidirectional insertion sorting unit uses the segmented sorted subsequence generated by the bitonic sorter as input, and records the maximum and minimum values of the subsequence. After all segmented subsequences are processed through the cascaded bidirectional insertion sorting units, a sorted sequence is obtained. The proposed architecture is implemented using the Verilog hardware description language (HDL) and synthesized using the Synopsys Design Compiler with a TSMC 90-nm cell library. The experimental results indicate that the proposed architecture can not only shorten sorting cycles but also reduce hardware area costs. Moreover, sorting cycles can be further shortened by increasing the parallelism of the proposed architecture. Under the configuration that 2048 32-bit data to be sorted and 16 data have to be processed simultaneously, the proposed architecture can improve the throughput-to-gate-count ratio by 16%, and throughput-to-power-consumption-ratio by 25% compared to the existing sorting design. The proposed architecture makes the most efficient use of hardware resources.
- Open Access
TMSR: Tiny Multi-path CNNs for Super Resolution
In this paper, we proposed a tiny multi-path CNN-based Super-Resolution (SR) method, called TMSR. We mainly refer to some tiny CNN-based SR methods, under 5k parameters. The main contribution of the proposed method is the improved multi-path learning and self-defined activated function. The experimental results show that TMSR obtains competitive image quality (i.e. PSNR and SSIM) compared to the related works under 5k parameters.
Open AccessTMSR: Tiny Multi-path CNNs for Super Resolution
In this paper, we proposed a tiny multi-path CNN-based Super-Resolution (SR) method, called TMSR. We mainly refer to some tiny CNN-based SR methods, under 5k parameters. The main contribution of the proposed method is the improved multi-path learning and self-defined activated function. The experimental results show that TMSR obtains competitive image quality (i.e. PSNR and SSIM) compared to the related works under 5k parameters.