A5086892451- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Image Retrieval and Classification Techniques
- Remote-Sensing Image Classification
- Advanced Measurement and Detection Methods
- Image Enhancement Techniques
- Advanced Algorithms and Applications
- Image and Video Quality Assessment
- Advanced Image and Video Retrieval Techniques
- Data Mining Algorithms and Applications
- Rough Sets and Fuzzy Logic
- Digital Media Forensic Detection
- Neurotransmitter Receptor Influence on Behavior
- Autophagy in Disease and Therapy
- Ergonomics and Musculoskeletal Disorders
- Advanced Sensor and Control Systems
- Remote Sensing and Land Use
- Elevator Systems and Control
- Energy and Environmental Systems
- Photoacoustic and Ultrasonic Imaging
- Color Science and Applications
- Schizophrenia research and treatment
- Industrial Vision Systems and Defect Detection
- Health and Well-being Studies
- Speech and Audio Processing
Nanchang University
2017-2023
Guizhou Provincial People's Hospital
2023
High-resolution remote sensing (HRRS) images contain abundant and complex visual contents. It is very important to extract powerful features represent the contents of HRRS in image retrieval. This letter proposes a region-based cascade pooling (RBCP) method aggregate convolutional from both pre-trained fine-tuned neural networks (CNNs). The RBCP adopts small regions, first uses max-pooling on feature maps last layer, then employs average-pooling max-pooled maps. Furthermore, map size related...
In this letter, a novel multiple image-based Gaussian noise level estimation (NLE) algorithm for natural images by jointly exploiting the level-aware feature extraction and local means (LM) techniques was proposed. We employed some efficient powerful features in form of vector to characterize levels across image contents. Based on this, we adopted LM scheme estimate an be estimated comparing similar preconstructed sample database. had verified accuracy efficiency proposed NLE large from...
In this letter, a two-stage noise level estimation (NLE) algorithm that jointly exploited automatic feature extraction and mapping model was proposed. contrast to existing NLE algorithms using hand-crafted features, we first utilized convolutional neural network-based automatically extract the level-aware features (NLAFs) in form of vector characterize distortion degree noisy image, i.e., level. Then, NLAF directly mapped its corresponding via pretrained model, obtaining fast reliable...
A number of previous studies have demonstrated the pivotal role PI3K/AKT signalling in cigarette smoke (CS)-induced emphysema, where phosphoinositide dependent protein kinase 1 (PDK1) is a critical component this pathway. Therefore, present study aimed to investigate effects PDK1 inhibitor (GSK-2334470) on expression levels PI3K, AKT, cyclin-dependent 2A (p16) and LC3B CS + extract (CSE)-induced mouse emphysema model. exposure intraperitoneal injections CSE were combined for 4 weeks...
To improve the evaluation accuracy of distorted images with various distortion types, an effective blind image quality assessment (BIQA) algorithm based on multi-window method and HSV color space is proposed in this paper. We generate multiple normalized feature maps (NFMs) by using to better characterize degradation from receptive fields different sizes. Specifically, distribution statistics are first extracted NFMs. Then, Pearson linear correlation coefficients between spatially adjacent...
Supervised image denoising methods based on deep neural networks require a large amount of noisy-clean or noisy pairs for network training. Thus, their performance drops drastically when the given is significantly different from training data. Recently, several unsupervised learning models have been proposed to reduce dependence Although only images learning, effect relatively weak compared with supervised methods. This paper proposes two-stage framework prior (DIP) enhance performance....
Purpose / Context - Residential energy consumption except heating (RECEH) plays an important role in China. The purposes of this study were to analyze RECEH Beijing and find out the influence factors RECEH. Methodology Approach A survey 2024 households was undertaken. investigat-ed residences located 23 residential communities two farmers markets. Questionnaire contents included six aspects: building information, envelope, living style, auxiliary household appliance, electricity other...
目的 现有方法存在特征提取时间过长、非对称失真图像预测准确性不高的问题，同时少有工作对非对称失真与对称失真立体图像的分类进行研究，为此提出了基于双目竞争的非对称失真立体图像质量评价方法。方法 依据双目竞争的视觉现象，利用非对称失真立体图像两个视点的图像质量衰减程度的不同，生成单目图像特征的融合系数，融合从左右视点图像中提取的灰度空间特征与HSV （hue-saturation-value）彩色空间特征。同时，量化两个视点图像在结构、信息量和质量衰减程度等多方面的差异，获得双目差异特征。并且将双目融合特征与双目差异特征级联为一个描述能力更强的立体图像质量感知特征向量，训练基于支持向量回归的特征—质量映射模型。此外，还利用双目差异特征训练基于支持向量分类模型的对称失真与非对称失真立体图像分类模型。结果 本文提出的质量预测模型在4个数据库上的SROCC （Spearman rank order correlation coefficient）和PLCC （Pearson linear coefficient）均达到0.95以上，在3个非对称失真数据库上的均方根误差（root of...