检索结果-内蒙古大学图书馆

Unsupervised Low-Light image Enhancement Based on Curve Estimation and Illumination Perception

signal, image and Video processing 2025年第8期19卷

作者： Xu, Xin Du, Jun School of Physics and Electronics Shandong Normal University Jinan250358 China

Low-light image enhancement is crucial for human vision and computer vision task, attracting significant attention. However, most current enhancement methods are supervised and lack the ability to adjust based on lighting conditions adaptively. Therefore, this paper proposes an end-to-end low-light image-enhancing method, which does not require paired datasets with ground truth. Specifically, we propose a lightweight deep neural network that estimates curve parameters and uses a set of no-reference loss functions to assess the quality of the enhanced image. Additionally, we design an illumination estimation module to adjust the enhancement parameters of low-light images adaptively. Experiments conducted on multiple datasets validate the effectiveness of our method, provide both qualitative and quantitative evaluations, and demonstrate its significant brightness enhancement capabilities. Our approach exhibits strong generalization abilities while preserving essential details critical for image interpretation. Rigorous experimental validation demonstrates that our method improves low-light images adaptively, enhancing both human visual perception and performance in the representative computer vision task of object detection. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Photointerpretation

来源：评论

学校读者我要写书评

暂无评论

NEAR-INFRARED image GUIDED neural NETWORKS FOR COLOR image DENOISING 44

NEAR-INFRARED IMAGE GUIDED NEURAL NETWORKS FOR COLOR IMAGE D...

引用

44th IEEE International Conference on Acoustics, Speech and signal processing (ICASSP)

作者： Wang, Xuehui Dai, Feng Ma, Yike Guo, Junbo Zhao, Qiang Zhang, Yongdong Chinese Acad Sci Inst Comp Technol Key Lab Intelligent Informat Proc Beijing Peoples R China Univ Chinese Acad Sci Beijing Peoples R China

ISBN: (纸本)9781479981311

Noisy color image and guided near-infrared (NIR) image can be jointly employed to eliminate noise and enhance details. Existing methods mostly rely on explicit designed filters and hand-crafted objective function optimization. These methods usually introduce erroneous structures from guidance signal. Besides, they are time-consuming and not suitable for real time applications. In this paper, we come up with a learning based method. The noisy color image and NIR image are fused, then fed into a fully convolutional neural network. The network learns a directly map from degraded image to restored sharp image. Our architecture can effectively eliminate image noise and transfer detail structure from guided image. Our trained network accepts any resolution of input image and runs in constant time. We evaluate the presented approach on both synthetic and real images. Results show that our approach outperforms the state-of-art methods.

关键词： Denoise color image MR image convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Impulse noise removal using residual convolutional neural networks

引用

AIP Conference Proceedings 2023年第1期2901卷

作者： A. K. C. Varma M. Dileep B. Prudhvi Raj G. Prasanna Kumar Vishnu Institute of Technology Bhimavaram Andhra Pradesh India

image denoising is one key concept in image restoration and it is widely used in various image processing applications. There are many traditional methods for image denoising existing, all these methods are based on filtering in spatial domain and frequency domain. This work focuses on removing of impulse noise, which converts the pixel values to zero or maximum. Proposed method comprises the convolutional neural network (CNN)to remove the impulse noise. The Residual CNN(RCNN) is used in the proposed method. The structure of the network consists of three stages that is convolutional layers followed by the residual block and finally convolutional layers. The skip connections in RCNN reduce the gradient vanish problem in traditional CNN based denoising methods. The proposed network trained by using dataset of 12 images. The stochastic gradient descent momentum (SGDM) optimizer used to optimize the weights. The RCNN trained using SGDM optimizer takes less time for convergence into minimum. The proposed network is tested with various testing images. The proposed RCNN based image denoising gives better results than the traditional median filter-based image denoising with respect to PSNR and SSIM.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fusion global and local deep representations with neural attention for aesthetic quality assessment

引用

signal processing-image COMMUNICATION 2019年 78卷 42-50页

作者： Zhang, Xiaodan Gao, Xinbo Lu, Wen Yu, Ying He, Lihuo Xidian Univ State Key Lab Integrated Serv Networks Sch Elect Engn Xian 710071 Shaanxi Peoples R China

In recent years, deep-learning based aesthetics assessment methods have shown promising results. However, existing methods can only achieve limited success because 1) most of the methods take one fixed-size patch as the training example, which loses the fine grained details and the holistic layout information, and 2) most of the methods ignore ordinal issues in image aesthetic assessment, ie. image scored 5.3 is more likely to be in the high quality class than image scored 4.5. To address these challenges, we presents a novel convolutional networks with two branches to encode global and local features. The first branch not only captures the spatial layout information but also feedbacks the top-down neural attention. The second branch selects the important attended region to extract the fine details features. A sobel-based attention layer is integrated with the second branch to enhance fine details encoding. Regarding the second problem, we combine the strength of classification approach and regression approach by a multi-task learning framework. Extensive experiments on challenging Aesthetic and Visual Analysis (AVA) dataset and *** dataset indicate the effectiveness of the proposed method.

关键词： image quality assessment image aesthetics analysis Deep neural network

来源：评论

学校读者我要写书评

暂无评论

Transformer-Based Speech Synthesizer Attribution in an Open Set Scenario

Transformer-Based Speech Synthesizer Attribution in an Open ...

引用

International Conference on Machine Learning and Applications (ICMLA)

作者： Emily R. Bartusiak Edward J. Delp Video and Image Processing Lab School of Electrical and Computer Engineering Purdue University West Lafayette IN

Speech synthesis methods can create realistic-sounding speech, which may be used for fraud, spoofing, and mis-information campaigns. Forensic methods that detect synthesized speech are important for protection against such attacks. Forensic attribution methods provide even more information about the nature of synthesized speech signals because they identify the specific speech synthesis method (i.e., speech synthesizer) used to create a speech signal. Due to the increasing number of realistic-sounding speech synthesizers, we propose a speech attribution method that generalizes to new synthesizers not seen during training. To do so, we investigate speech synthesizer attribution in both a closed set scenario and an open set scenario. In other words, we consider some speech synthesizers to be "known" synthesizers (i.e., part of the closed set) and others to be "unknown" synthesizers (i.e., part of the open set). We represent speech signals as spectrograms and train our proposed method, known as compact attribution transformer (CAT), on the closed set for multi-class classification. Then, we extend our analysis to the open set to attribute synthesized speech signals to both known and unknown synthesizers. We utilize a t-distributed stochastic neighbor embedding (tSNE) on the latent space of the trained CAT to differentiate between each unknown synthesizer. Additionally, we explore poly-1 loss formulations to improve attribution results. Our proposed approach successfully attributes synthesized speech signals to their respective speech synthesizers in both closed and open set scenarios.

关键词： Training Synthesizers Forensics Machine learning Transformers Fraud Speech synthesis

来源：评论

学校读者我要写书评

暂无评论

Single-image Reflection Removal via a Two-Stage Background Recovery Process

引用

IEEE signal processing LETTERS 2019年第8期26卷 1237-1241页

作者： Li, Tinglian Lun, Daniel P. K. Hong Kong Polytech Univ Dept Elect & Informat Engn Hong Kong Peoples R China

The reflection problem often occurs when imaging through a semitransparent material such as glass. It degrades the image quality and affects the subsequent analyses on the image. Traditional single-image based reflection removal methods assume the reflection is blurry. Deep neural networks (DNNs) are, then, used to identify the blurry reflection and remove it. However, it is often that the blurry reflection still contains strong edges. They will be treated as the background and kept in the image. In this letter, we propose a novel two-stage DNN based reflection removal algorithm. In the first stage, we include a new feature reduction term in the loss function when training the network. Due to its strong reflection suppression ability, the reflection components in the image can he more effectively suppressed. However, it will also attenuate the gradient values of the background image. For recovering the background, in the second stage, we first estimate a reflection gradient confidence map based on the initial estimation result and use it to identify the strong background gradients. Then, we use a generative adversarial network to reconstruct the background image from its gradients. Experimental results show that the proposed two-stage approach can give a superior performance compared with the state-of-the-art DNN based methods.

关键词： image reflection removal blind image separation deep neural network

来源：评论

学校读者我要写书评

暂无评论

Deep learning model for real-time image compression in Internet of Underwater Things (IoUT)

Deep learning model for real-time image compression in Inter...

引用

作者： Krishnaraj, N. Elhoseny, Mohamed Thenmozhi, M. Selim, Mahmoud M. Shankar, K. Department of Computer Science and Engineering SASI Institute of Technology and Engineering TadepalligudemAndhra Pradesh India Faculty of Computers and Information Mansoura University Mansoura Egypt Department of IT SRM Institute of Science and Technology KanchipuramTamil Nadu India Department of Mathematics Al-Aflaj College of Science and Human Studies Prince Sattam Bin Abdulaziz University Al Kharj Saudi Arabia School of Computing Kalasalingam Academy of Research and Education Krishnankoil India

Recently, the advancements of Internet-of-Things (IoT) have expanded its application in underwater environment which leads to the development of a new field of Internet of Underwater Things (IoUT). It offers a broader view of applications such as atmosphere observation, habitat monitoring of sea animals, defense and disaster prediction. Data transmission of images captured by the smart underwater objects is very challenging due to the nature of underwater environment and necessitates an efficient image transmission strategy for IoUT. In this paper, we model and implement a discrete wavelet transform (DWT) based deep learning model for image compression in IoUT. For achieving effective compression with better reconstruction image quality, convolution neural network (CNN) is used at the encoding as well as decoding side. We validate DWT–CNN model using extensive set of experimentations and depict that the presented deep learning model is superior to existing methods such as super-resolution convolutional neural networks (SRCNN), JPEG and JPEG2000 in terms of compression performance as well as reconstructed image quality. The DWT–CNN model attains an average peak signal-to-noise ratio (PSNR) of 53.961 with average space saving (SS) of 79.7038%. © 2019, Springer-Verlag GmbH Germany, part of Springer Nature.

关键词： image compression

来源：评论

学校读者我要写书评

暂无评论

Deep CNNs as a method to classify rotating objects based on monostatic RCS

引用

IET RADAR SONAR AND NAVIGATION 2019年第7期13卷 1092-1100页

作者： Wengrowski, Eric Purri, Matthew Dana, Kristin Huston, Andrew Rutgers State Univ Dept Elect & Comp Engn New Brunswick NJ 08903 USA Lockheed Martin Radar Grp Moorestown NJ USA

Radar systems emit a time-varying signal and measure the response of a radar-reflecting surface. In the case of narrowband, monostatic radar signal domain, all spatial information is projected into a radar cross-section (RCS) scalar. The authors address the challenging problem of determining shape class using monostatic RCS estimates collected as a time series from a rotating object tumbling with unknown motion parameters under detectability limitations and signal noise. Previous shape classification methods have relied on image-like synthetic aperture radar or multistatic (multiview) radar configurations with known geometry. Convolutional neural networks (CNNs) have revolutionised learning tasks in the computer vision domain by leveraging images and video rich with high-resolution two-dimensional (2D) or 3D spatial information. They show that a feed-forward CNN can be trained to successfully classify object shape using only noisy monostatic RCS signals with unknown motion. They construct datasets containing over 100,000 simulated RCS signals belonging to different shape classes. They introduce deep neural network architectures that produce 2% classification error on testing data. They also introduce a refinement network that transforms simulated signals to appear more realistic and improve training utility. The results are a pioneering step toward the recognition of more complex targets using narrowband, monostatic radar.

关键词： object detection learning (artificial intelligence) radar signal processing object recognition radar cross-sections computer vision radar imaging radar detection image classification feature extraction synthetic aperture radar neural nets spatial information object shape noisy monostatic RCS signals 100 simulated RCS signals 000 simulated RCS signals different shape classes deep neural network architectures 2% classification error narrowband CNNs radar systems time-varying signal radar-reflecting surface monostatic radar signal domain radar cross-section scalar shape class monostatic RCS estimates time series rotating object unknown motion parameters detectability limitations signal noise previous shape classification methods image-like synthetic aperture radar multistatic radar configurations convolutional neural networks computer vision domain high-resolution two-dimensional

来源：评论

学校读者我要写书评

暂无评论

image Based Tumor Cells Identification Using Convolutional neural Network and Auto Encoders

引用

TRAITEMENT DU signal 2019年第5期36卷 445-453页

作者： Wajeed, Mohammed Abdul Sreenivasulu, Vallamchetty Keshav Mem Inst Technol Dept Comp Sci & Engn Hyderabad 500029 India

The convolutional neural network (CNN) and other neural networks (NNs) provide promising tools for robotized characterization of tumor cells. However, the tumor growth areas in ultrasound images are normally obscure, with uncertain edges. It is not acceptable to prepare ultrasound images straightforwardly with the CNN. To solve the problem, this paper puts forward a faster region-convolutional neural network (R-CNN) to identify tumor cells with the aid of auto encoders Taking two fully-connected layers with dropout and ReLU enactments as the base, the proposed faster R-CNN adopts 3D convolutional and max pooling layers, enabling the user to extract features from potential tumor growth areas. In addition, the thin and deep layers of the network were connected to facilitate the identification of blurry or small tumor growth areas. Experimental results show that the proposed faster R-CNN with auto encoders outperformed traditional data mining and artificial intelligence (AI) methods in prediction accuracy of tumor cells.

关键词： convolutional neural network region-convolutional neural network tumor cells pre processing clustering classification tumor prediction

来源：评论

学校读者我要写书评

暂无评论

Enhancing meibography based assessment of gland morphology by utilizing an image-rotating Mask R-CNN approach

引用

Biomedical signal processing and Control 2025年 109卷

作者： Paściak, Agnieszka Piwowarczyk, Patrycja K. Iskander, D. Robert Szczęsna-Iskander, Dorota H. Department of Optics and Photonics Wroclaw University of Science and Technology Wybrzeze Wyspianskiego 27 Wroclaw50-370 Poland Department of Biomedical Engineering Wroclaw University of Science and Technology Wybrzeze Wyspianskiego 27 Wroclaw50-370 Poland

Accurate analysis of meibomian gland morphology based on meibography images is of great importance for the diagnosis of dry eye disease. However, it is still a difficult task due to the time-consuming and variability of traditional methods. Nowadays, machine learning-based approaches are becoming increasingly common. Nevertheless, there is currently no universal model for evaluating images from different devices, and the segmentation quality is still insufficient. In this study, a novel approach is used that combines the DeepLabV3 neural network for tarsus segmentation with an image-rotating implementation of the Mask R-CNN neural network for segmenting meibomian gland instances. The proposed method achieves an average Jaccard index of 66.3% and an intersection over union of 70.5% on the test dataset, outperforming other methods while demonstrating competitive performance for the two leading meibography devices. Furthermore, the proposed approach enables precise determination of glandular morphological parameters, including length, thickness, tortuosity, and glands area ratio. These findings highlight the diagnostic value of Mask R-CNN in providing effective and standardised assessments of meibomian gland morphology. © 2025 The Authors

关键词： image segmentation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：