Zeqi Yu

Dr. Zeqi Yu


Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 May 2024 in Information
Verified
Journal Article
Open access
PDF file
BJBin Jiang
Bin Jiang
NLNanxing Li
Nanxing Li
XCXiaomei Cui
Xiaomei Cui
Weihua LiuWeihua Liu
Weihua Liu
ZYZeqi Yu
Zeqi Yu
XYXie Yongheng
Xie Yongheng
To avoid the overfitting problem of the network model and improve the facial expression recognition effect of partially occluded facial images, an improved facial expression recognition algorithm based on MobileViT has been proposed. Firstly, in order to obtain features that are useful and richer for experiments, deep convolution operations are added to the inverted residual blocks of this network, thus improving the facial expression recognition rate. Then, in the process of dimension reduction, the activation function can significantly improve the convergence speed of the model, and then quickly reduce the loss error in the training process, as well as to preserve the effective facial expression features as much as possible and reduce the overfitting problem. Experimental results on RaFD, FER2013, and FER2013Plus show that this method has significant advantages over mainstream networks and the network achieves the highest recognition rate.

29 August 2014 in Circuits, Systems, and Signal Processing
Verified
Journal Article
Closed access
PDF file
ZYZeqi Yu
Zeqi Yu
YFYangyu Fan
Yangyu Fan
LSLongfei Shi
Longfei Shi
GLGuoyun Lv
Guoyun Lv

16 June 2020 in Journal of Power Electronics
Verified
Journal Article
Closed access
PDF file
ZYZeqi Yu
Zeqi Yu
GBGe Bai
Ge Bai
KZKe Zhang
Ke Zhang
JXJin Xu
Jin Xu

7 September 2019 in Circuits, Systems, and Signal Processing
Verified
Journal Article
Closed access
PDF file
ZYZeqi Yu
Zeqi Yu
EZEnguang Zhang
Enguang Zhang
XCXiaolei Chen
Xiaolei Chen
GBGe Bai
Ge Bai
JXJin Xu
Jin Xu
XWXu Wang
Xu Wang

11 September 2024 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
ZYZeqi Yu
Zeqi Yu
HLHaokai Liu
Haokai Liu
LLLei Liu
Lei Liu
NZNing Zhang
Ning Zhang
KOKaoru Ota
Kaoru Ota
MDMianxiong Dong
Mianxiong Dong
Filterless digital class-D amplifiers (CDAs) are very attractive for audio-visual devices thanks to their high efficiency, small size and low cost. However, their high-frequency switchingmode operation will cause electromagnetic interference (EMI) problems. Spread spectrum techniques are often used to solve the EMI problems but will cause baseband distortions in amplifiers. In this article, an all-digital spread spectrum method with distortion correction is proposed to make filterless digital CDAs achieve low EMI emissions and high signal to noise and distortion ratio (SNDR). This method mainly utilizes the random numbers generated by a pseudo-random number generator to randomize the pulse position and pulse repetition frequency (PRF) of the uniform-sampling pulse width modulation (UPWM) generator output signal to achieve the effect of spread spectrum. Moreover, the high open-loop gain characteristic of the digital sigma-delta modulator in the baseband and the state-space reconstruction technique are utilized to correct the baseband distortions caused by UPWM and spread spectrum. The feasibility and effectiveness of the proposed method are verified by simulation and experimental results.

24 October 2026 in ACM Transactions on Embedded Computing Systems
Verified
Journal Article
Open access
ZYZeqi Yu
Zeqi Yu
HZHeng Zhao
Heng Zhao
ZLZuhe Li
Zuhe Li
YPYushan Pan
Yushan Pan
Xinfei GuoXinfei Guo
Xinfei Guo
Contrast-Limited Adaptive Histogram Equalization (CLAHE) is widely used in industrial and embedded imaging, yet a fixed Clip Limit often fails to balance detail enhancement and noise suppression across diverse scenes while maintaining stall-free, one-pixel-per-cycle pixel-stream throughput. This article proposes a deployment-driven, perception-guided software–hardware co-design for CLAHE, where perceptual guidance from offline no-reference image quality assessment (NR-IQA) signals is distilled into a low-dimensional policy that only consumes hardware-observable statistics, and the field-programmable gate array (FPGA) architecture embeds closed-loop control without stalling the main pixel path. On the learning side (offline, in software), Clip Limit tuning is formulated as a Markov decision process (MDP) and learned with Soft Actor-Critic (SAC) using a dual-teacher distillation scheme that transfers privileged NR-IQA guidance to a low-dimensional student policy. Posterior trajectory selection (PTS) and Q4.12 quantization-aware distillation further improve robustness under partial observability and fixed-point inference. On the hardware side, we implement an FPGA single-clock-domain streaming pipeline that integrates feature extraction, student inference, and CLAHE mapping while preserving one-pixel-per-cycle throughput, and we reduce critical-path complexity via division elimination and separable fixed-point interpolation. The controller performs one update per frame and commits the new Clip Limit at the frame boundary; therefore, a five-step horizon corresponds to a five-frame gradual adaptation with intra-frame consistency. Experiments show that the proposed controller is competitive with strong fixed-parameter baselines in mean NR-IQA scores, while reducing failure-sensitive upper-tail degradation using hardware-observable inputs only.