Peng Feng

Peng Feng


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
5 July 2024 in Sensors
Verified
Journal Article
Open access
PDF file
ZWZhe Wang
Zhe Wang
XYXu Yang
Xu Yang
NTNa Tian
Na Tian
MLMin Liu
Min Liu
ZCZiteng Cai
Ziteng Cai
LLLIYUAN LIU
LIYUAN LIU
Low-light imaging capabilities are in urgent demand in many fields, such as security surveillance, night-time autonomous driving, wilderness rescue, and environmental monitoring. The excellent performance of SPAD devices gives them significant potential for applications in low-light imaging. This article presents a 64 (rows) × 128 (columns) SPAD image sensor designed for low-light imaging. The chip utilizes a three-dimensional stacking architecture and microlens technology, combined with compact gated pixel circuits designed with thick-gate MOS transistors, which further enhance the SPAD’s photosensitivity. The configurable digital control circuit allows for the adjustment of exposure time, enabling the sensor to adapt to different lighting conditions. The chip exhibits very low dark noise levels, with an average DCR of 41.5 cps at 2.4 V excess bias voltage. Additionally, it employs a denoising algorithm specifically developed for the SPAD image sensor, achieving two-dimensional grayscale imaging under 6 × 10−4 lux illumination conditions, demonstrating excellent low-light imaging capabilities. The chip designed in this paper fully leverages the performance advantages of SPAD image sensors and holds promise for applications in various fields requiring low-light imaging capabilities.

3 March 2025 in Chips
Verified
Journal Article
Open access
PDF file
JCJingyang Chen
Jingyang Chen
NCNanbo Chen
Nanbo Chen
ZWZhe Wang
Zhe Wang
RDRUNJIANG Dou
RUNJIANG Dou
JLJian Liu
Jian Liu
GWGang Wang
Gang Wang
High-dynamic-range (HDR) technology enhances the capture of luminance beyond the limits of traditional images, facilitating the capture of more nuanced and lifelike visual effects. This advancement has profound implications across various sectors, such as medical imaging, augmented reality (AR), virtual reality (VR), and autonomous driving systems. The evolution of complementary metal-oxide semiconductor (CMOS) image sensor (CIS) manufacturing techniques, particularly through backside illumination (BSI) and advancements in three-dimensional (3D) stacking architectures, is driving progress in HDR’s capabilities. This paper provides a review of the technologies developed over the past six years that augment the dynamic range (DR) of CIS. It systematically introduces and summarizes the implementation methodologies and distinguishing features of each technology.

4 December 2025 in Remote Sensing
Verified
Journal Article
Open access
PDF file
DZDanshu Zhou
Danshu Zhou
YXYushan Xiong
Yushan Xiong
SYShuangming Yu
Shuangming Yu
PFPeng Feng
Peng Feng
JLJian Liu
Jian Liu
LLLIYUAN LIU
LIYUAN LIU
Ship detection in remote sensing imagery is hindered by cluttered backgrounds, large variations in scale, and random orientations, limiting the performance of detectors designed for natural images. We propose RSDB-Net, a Rotation-Sensitive Dual-Branch Detection Network that introduces innovations in feature extraction, fusion, and detection. The Swin Transformer–CNN Backbone (STCBackbone) combines a Swin Transformer for global semantics with a CNN branch for local spatial detail, while the Feature Conversion and Coupling Module (FCCM) aligns and fuses heterogeneous features to handle multi-scale objects, and a Rotation-sensitive Cross-branch Fusion Head (RCFHead) enables bidirectional interaction between classification and localization, improving detection of randomly oriented targets. Additionally, an enhanced Feature Pyramid Network (eFPN) with learnable transposed convolutions restores semantic information while maintaining spatial alignment. Experiments on DOTA-v1.0 and HRSC2016 show that RSDB-Net performs better than the state of the art (SOTA), with mAP-ship values of 89.13% and 90.10% (+5.54% and +44.40% over the baseline, respectively), and reaches 72 FPS on an RTX 3090. RSDB-Net also demonstrates strong generalization and scalability, providing an effective solution for rotation-aware ship detection.

1 September 2025 in Journal of Semiconductors
Verified
Journal Article
Closed access
HZHuanhui Zhang
Huanhui Zhang
CZChi Zhang
Chi Zhang
XYXu Yang
Xu Yang
ZWZhe Wang
Zhe Wang
CSCong Shi
Cong Shi
LLLiyuan Liu
Liyuan Liu
The event-based vision sensor (EVS), which can generate efficient spiking data streams by exclusively detecting motion, exemplifies neuromorphic vision methodologies. Generally, its inherent lack of texture features limits effectiveness in complex vision processing tasks, necessitating supplementary visual information. However, to date, no event-based hybrid vision solution has been developed that preserves the characteristics of complete spike data streams to support synchronous computation architectures based on spiking neural network (SNN). In this paper, we present a novel spike-based sensor with digitized pixels, which integrates the event detection structure with the pulse frequency modulation (PFM) circuit. This design enables the simultaneous output of spiking data that encodes both temporal changes and texture information. Fabricated in 180 nm process, the proposed sensor achieves a resolution of 128 × 128, a maximum event rate of 960 Meps, a grayscale frame rate of 117.1 kfps, and a measured power consumption of 60.1 mW, which is suited for high-speed, low-latency, edge SNN-based vision computing systems.

1 January 2019 in IEICE Electronics Express
Verified
Journal Article
Closed access
PDF file
QDQian Di
Qian Di
ZZZhongxing Zhang
Zhongxing Zhang
HLHonglong Li
Honglong Li
ZZZhao Zhang
Zhao Zhang
PFPeng Feng
Peng Feng
NWNanjian Wu
Nanjian Wu
This paper proposes novel single event upset (SEU) failure probability evaluation and periodic scrubbing techniques for hierarchical parallel vision processors. To automatically evaluate the SEU failure probability and identify all the critical elements in a processor, complementary fault injection methods based on logic circuit simulator and Perl script are proposed. These methods can be used to randomly inject faults into D flip-flops (DFFs) and various types of memory at the register transfer level (RTL) as well as to evaluate the vision processor performance. Based on the evaluation results, an accurate periodic scrubbing technique is proposed to increase the processor availability. The results denote that the peak availability of the processor over a period of one year can be improved from 18% to 99.9% after scrubbing the RISC program memory for a period of 104 s. Therefore, we can improve the fault-tolerance performance of a vision processor while avoiding unnecessary area and power costs using techniques ranging from evaluation to mitigation.

27 December 2019 in arXiv
Verified
Preprint
Open access
PDF file
MZMingxin Zhao
Mingxin Zhao
LCLi Cheng
Li Cheng
XYXu Yang
Xu Yang
PFPeng Feng
Peng Feng
LLLIYUAN LIU
LIYUAN LIU
NWNANJIAN WU
NANJIAN WU
Infrared small target detection is a key technique in infrared search and tracking (IRST) systems. Although deep learning has been widely used in the vision tasks of visible light images recently, it is rarely used in infrared small target detection due to the difficulty in learning small target features. In this paper, we propose a novel lightweight convolutional neural network TBC-Net for infrared small target detection. The TBCNet consists of a target extraction module (TEM) and a semantic constraint module (SCM), which are used to extract small targets from infrared images and to classify the extracted target images during the training, respectively. Meanwhile, we propose a joint loss function and a training method. The SCM imposes a semantic constraint on TEM by combining the high-level classification task and solve the problem of the difficulty to learn features caused by class imbalance problem. During the training, the targets are extracted from the input image and then be classified by SCM. During the inference, only the TEM is used to detect the small targets. We also propose a data synthesis method to generate training data. The experimental results show that compared with the traditional methods, TBC-Net can better reduce the false alarm caused by complicated background, the proposed network structure and joint loss have a significant improvement on small target feature learning. Besides, TBC-Net can achieve real-time detection on the NVIDIA Jetson AGX Xavier development board, which is suitable for applications such as field research with drones equipped with infrared sensors.