A5101475406- Advanced Neural Network Applications
- Error Correcting Code Techniques
- Advanced Wireless Communication Techniques
- Advanced Memory and Neural Computing
- Neural Networks and Applications
- Antenna Design and Analysis
- Higher Education and Teaching Methods
- Cooperative Communication and Network Coding
- Quantum Computing Algorithms and Architecture
- Quantum-Dot Cellular Automata
- Neural Networks Stability and Synchronization
- Microwave Engineering and Waveguides
- Advanced Image and Video Retrieval Techniques
- RFID technology advancements
- Radiation Detection and Scintillator Technologies
- Blockchain Technology Applications and Security
- Cryptography and Data Security
- Particle Detector Development and Performance
- Quantum Information and Cryptography
- Advanced Decision-Making Techniques
- Medical Image Segmentation Techniques
- Advanced Antenna and Metasurface Technologies
- Privacy-Preserving Technologies in Data
- Ferroelectric and Negative Capacitance Devices
- Education and Work Dynamics
University of Notre Dame
1997-2025
Northwestern Polytechnical University
2017-2025
State Key Laboratory of Quantum Optics and Quantum Optics Devices
2022-2023
Shanxi University
2022-2023
Qingdao Academy of Intelligent Industries
2020
China Shipbuilding Industry Corporation (China)
2019
Hong Kong Polytechnic University
2014-2016
Jiangsu Maritime Institute
2015
Qufu Normal University
2013
Foshan University
2008-2013
With pervasive applications of medical imaging in health-care, biomedical image segmentation plays a central role quantitative analysis, clinical diagnosis, and intervention. Since manual annotation suffers limited reproducibility, arduous efforts, excessive time, automatic is desired to process increasingly larger scale histopathological data. Recently, deep neural networks (DNNs), particularly fully convolutional (FCNs), have been widely applied segmentation, attaining much improved...
In the recent past, success of Neural Architecture Search (NAS) has enabled researchers to broadly explore design space using learning-based methods. Apart from finding better neural network architectures, idea automation also inspired improve their implementations on hardware. While some practices hardware machine-learning have achieved remarkable performance, traditional concept is still followed: a architecture first structured with excellent test accuracy, and then compressed optimized...
Based on the passband characteristics of frequency selective surface (FSS), a wideband transmitarray (TA) with linear polarization rotation at W-band is proposed. The middle layer unit cell pattern arrow-shaped structure, which can rotate incoming wave by 90°. A pair parallel strip polarizers employed to enhance conversion efficiency. prototype <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$25\times 25$...
Abstract Quantum key distribution (QKD) can provide information-theoretically secure keys for two parties of legitimate communication, and information reconciliation, as an indispensable component QKD systems, correct errors present in raw based on error-correcting codes. In this paper, we first describe the basic knowledge reconciliation its impact continuous variable QKD. Then introduce schemes corresponding error correction codes employed. Next, rate-compatible codes, hardware...
In this paper, we propose a new class of quasi-cyclic low-density parity-check (QC-LDPC) codes, namely cyclically-coupled QC-LDPC (CC-QC-LDPC) and their RAM-based decoder architecture. CC-QC-LDPC codes have simple structure are constructed by cyclically-coupling number subcodes. They can achieve throughput error performance as excellent LDPC convolutional but with much lower hardware requirements. therefore promising candidates for future generations communication systems such long-haul...
With pervasive applications of medical imaging in health-care, biomedical image segmentation plays a central role quantitative analysis, clinical diagno- sis, and intervention. Since manual anno- tation su ers limited reproducibility, arduous e orts, excessive time, automatic is desired to process increasingly larger scale histopathological data. Recently, deep neural networks (DNNs), par- ticularly fully convolutional (FCNs), have been widely applied segmenta- tion, attaining much improved...
A wideband planar dipole array based on artificial magnetic conductor (AMC) surface is presented for phased applications. One in-phase hexagonal AMC cell realized firstly. An X-band modified by employing a detached H-shaped patch and four L-shaped stubs impedance bandwidth enhancement. The active VSWR with different scanning angles the radiation pattern of center element in small are simulated. For demonstration, prototype 45° linearly polarized elements 4 × 32 fabricated measured. From...
The computation of the reliability inferences among variables a single-parity-check (SPC) code is common challenge to implementation channel decoders. Applicable variety computational mechanisms using disparate kernels, parallel-routing network has been developed, which is, compared state-of-art structure, exploring best its parallel nature for an improved timing performance. Furthermore, we assure that proposed structure no degradation in neither accuracy nor hardware complexity. With this...
The rapid development of Internet Things is yielding a huge volume time series data, the real-time mining which becomes major load for data centers. computation bottleneck in distance measure, dynamic warping (DTW) one most widely used measures. Recently, various software optimization and hardware acceleration techniques have been proposed DTW acceleration. However, throughput energy efficiency are still big concerns considering ever-increasing times series. In this paper, we propose...
Cellular neural networks (CeNNs) have been widely adopted in image processing tasks. Recently, various hardware implementations of CeNNs emerged the literature, with Field Programmable Gate Array (FPGA) being one most popular choices due to its high flexibility and low time-to-market. However, typically involve extensive computations a recursive manner. As an example, simply process 1,920 × 1,080 pixels requires 4--8 Giga floating point multiplications (for 3 templates 50–100 iterations),...
With the need for increased care and welfare of rapidly aging population, mobile telemedicine is becoming popular providing remote health to increase quality life. Recently, image analysis being actively applied medical diagnosis treatment, in which segmentation fundamental importance other processing such as visualization detection. However, given tasks challenges transmitting large volume high-resolution images real-time constraints that are commonly present telemedicine, best done at...
Deep neural networks have demonstrated their great potential in recent years, exceeding the performance of human experts a wide range applications. Due to large sizes, however, compression techniques such as weight quantization and pruning are usually applied before they can be accommodated on edge. It is generally believed that leads degradation, plenty existing works explored strategies aiming at minimum accuracy loss. In this paper, we argue quantization, which essentially imposes...
Due to the fast growing industry of smart cars and autonomous driving, advanced driver assistance systems (ADAS) with its applications have attracted a lot attention. As crucial part ADAS, obstacle detection has been challenge due real-tme resource-constraint requirements. Cellular neural network (CeNN) popular for detection, however suffers from high computation complexity. In this paper we propose compressed CeNN framework real-time ADAS in embedded FPGAs. Particularly, parameter...
A multi-mode QC-LDPC decoder is proposed to satisfy the 802.11n/ac WiFi standard. With code-specific design, overall performance of enhanced while ensuring an on-the-fly reconfigurable ability. The architecture has been synthesized using FPGA for measurements. state-of-art error rate and implementation complexity are reported. Meanwhile, throughput increased range from 382 MHz 1852 MHz.
Due to increasing privacy concerns, neural network (NN) based secure inference (SI) schemes that simultaneously hide the client inputs and server models attract major research interests. While existing works focused on developing protocols for NN-based SI, in this work, we take a different approach. We propose NASS, an integrated framework search tailored NN architectures designed specifically SI. In particular, model cryptographic as design elements with associated reward functions. The...
With the need for increased care and welfare of rapidly aging population, mobile telemedicine is becoming popular providing remote health to increase quality life. Recently, image analysis being actively applied medical diagnosis treatment, in which segmentation fundamental importance other processing such as visualization detection. However, given tasks challenges transmitting large volume high-resolution images real-time constraints that are commonly present telemedicine, best done at...
Quantization has become an essential technique in compressing deep neural networks for deployment onto resource-constrained hardware. It is noticed that, the hardware efficiency of implementing quantized highly coupled with actual values to be into, and therefore, given bit widths, we can smartly choose a value space further boost efficiency. For example, using weights only integer powers two, multiplication fulfilled by operations. Under such circumstances, however, existing...
A digital additive white Gaussian noise (AWGN) generator has recently become a public focus with the increasing demand of hardware simulation in researches on communication discipline. Some successful ideas software, such as Box-Muller method, are then proposed and implemented into platforms architecture design attracted new studies. Then high-precision small-error generators appeared literatures from time to time. Considering that academic applications prefer higher throughput data...