A5100350192- Advanced Image Processing Techniques
- Image Processing Techniques and Applications
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Infrared Target Detection Methodologies
- Image Enhancement Techniques
- Optical Systems and Laser Technology
- Video Surveillance and Tracking Methods
- Advanced Optical Imaging Technologies
- Advanced Neural Network Applications
- Image and Video Stabilization
- Advanced Measurement and Detection Methods
- Advanced Image and Video Retrieval Techniques
- Visual Attention and Saliency Detection
- Optical Coherence Tomography Applications
- Photoacoustic and Ultrasonic Imaging
- Optical measurement and interference techniques
- Calibration and Measurement Techniques
- Advanced optical system design
- Remote-Sensing Image Classification
- Adaptive optics and wavefront sensing
- Sparse and Compressive Sensing Techniques
- Face and Expression Recognition
- Planetary Science and Exploration
Zhejiang University
2016-2025
Beijing Tongren Hospital
2025
Capital Medical University
2025
Southeast University
2006-2025
Industrial and Commercial Bank of China
2024
China Automotive Engineering Research Institute
2024
North China Electric Power University
2018-2024
Xi'an University of Technology
2023-2024
Changchun University of Science and Technology
2023-2024
University of Chinese Academy of Sciences
2024
Abstract Temporal solitons have been the focus of much research due to their fascinating physical properties. These can form bound states, which are fundamentally crucial modes in fiber laser and present striking analogies with matter molecules counterparts, means they potential applications large-capacity transmission all-optical information storage. Although traditionally, second-order dispersion has dominant for conventional solitons, recent experimental theoretical shown that...
Recent years have witnessed a dramatic increase in the quantity of image data collected, due to advances fields such as medical imaging, reconnaissance, surveillance, astronomy, multimedia etc. With this has come need be able store, transmit, and query large volumes efficiently. A common operation on databases is retrieval all images that are similar image. For this, database often represented vectors high-dimensional space answered by retrieving proximal space, under suitable similarity...
Color, as a notable and stable attribute of vehicles, can serve useful reliable cue in variety applications intelligent transportation systems. Therefore, vehicle color recognition natural scenes has become an important research topic this area. In paper, we propose deep-learning-based algorithm for automatic recognition. Different from conventional methods, which usually adopt manually designed features, the proposed is able to adaptively learn representation that more effective task...
Correcting the optical aberrations and manufacturing deviations of cameras is a challenging task. Due to limitation on volume demand for mass production, existing mobile terminals cannot rectify degradation. In this work, we systematically construct perturbed lens system model illustrate relationship between deviated parameters spatial frequency response measured from photographs. To further address issue, an optimization framework proposed based build proxy machining samples' SFRs. Engaging...
Abstract The purpose of human object detection is to obtain the number people and their position in images, which one core problems field machine vision. However, high missing rate from small- medium-sized bodies due large variety scale tasks still influences performance detection. To solve above problem, this paper proposed an improved ASPP_BiFPN_YOLOv4 (ABYOLOv4) method detect In detail, Atrous Spatial Pyramid Pooling (ASPP) module was used replace original increase receptive level network...
Low-light imaging with handheld mobile devices is a challenging issue. Limited by the existing models and training data, most methods cannot be effectively applied in real scenarios. In this paper, we propose new low-light image restoration method using complementary information of short- long-exposure images. We first novel data generation to synthesize realistic raw images simulating pipeline environment. Then, design long-short-exposure fusion network (LSFNet) deal problems fusion,...
As the popularity of mobile photography continues to grow, considerable effort is being invested in reconstruction degraded images. Due spatial variation optical aberrations, which cannot be avoided during lens design process, recent commercial cameras have shifted some these correction tasks from postprocessing systems. However, without engaging with parameters, systems only achieve limited for aberrations. In this work, we propose a practical method recovering degradation caused by...
Image dehazing is a typical low-level visual task. With the continuous improvement of network performance and introduction various prior knowledge, ability image becoming stronger. However, existing methods have problems such as inability to obtain real shooting datasets, unreliable processes, difficulty deal with complex lighting scenes. To solve these problems, we propose new haze model combining optical scattering computer graphics rendering. Based on model, high-quality widely applicable...
Resolving inverse heat conduction problems (IHCPs) based on interior measured temperatures is a reliable approach for aerodynamic prediction. In most IHCP studies, it generally assumed that the data obtained by temperature sensors are positional temperature, whereas in actual measurements, thermocouple undergoes transient transfer process, leading to slight bias its readings compared desired at target location. this paper, model established simulate sensor dynamic response effect. Then,...
Coaxial surface junction thermocouples (CSJTs) are commonly applied to transient thermal measurements. Based on the measurement principle of CSJT, one-dimensional analytical inversion is for obtaining heat fluxes. However, with thermophysical properties difference between wall material and transverse conduction occurs that leads a error temperature by CSJT. Hence, conventional data processing methods can lead flux error. Therefore, quantitative analysis lateral structural transfer should be...
Introduction: Lung cancer is a major health concern worldwide owing to its high incidence and mortality rates. Therefore, identification of new therapeutic targets strategies for lung critical improving patient outcomes. Peptidyl arginine deiminase 4 (PAD4) promotes tumor growth metastasis by catalyzing the citrullination histones, making it potential target. Although PAD4 inhibitors have shown in treatment variety tumors, existing lack sufficient specificity cause substantial systemic...