检索结果-内蒙古大学图书馆

Extraction method of microscopic moving particle basing on multimodal background model

ICIC Express Letters, Part B: Applications 2015年第12期6卷 3153-3158页

作者： Zhang, Yanbin Yang, Zhangzhen Zhang, Hui Qin, Lei Image Processing and Image Communication Lab Jiangsu Key Lab of Wireless Communications Nanjing University of Posts and Telecommunications No. 66 New Mofan Road Nanjing210003 China Jiangsu Key Lab of Wireless Communications Nanjing University of Posts and Telecommunications No. 66 New Mofan Road Nanjing210003 China

Target extraction is a key technology for image measurement of moving particles distributed in fluidic system. In this paper, we propose a novel moving particle extraction method based on multimodal characteristic of the image. For the proposed particle extraction method, we firstly analyze the origin of multimodal characteristic of image background in the microscopic observation system, and then we design multimodal background model for particle extraction. For improving the performance of the model, we exploit the moving principle in the micro channel, and discuss the setting of learning rate. The experimental results demonstrate that the proposed methods achieve good performance compared with several traditional methods. © 2015, ICIC Express Letters Office. All rights reserved.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Comparison of regressors on 3D visual discomfort prediction

Comparison of regressors on 3D visual discomfort prediction

引用

IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB)

作者： Runfeng Huang Jun Zhou Xiao Gu Ya Zhang Alan C. Bovik Shanghai Key Laboratory of Digital Media Processing and Transrnissions Shanghai China Institute of Image Communication and Network Engineering Shanghai Jiao Tong University Laboratory for Image and Video Engineering University of Texas at Austin Austin USA

ISBN: (纸本)9781467390453

Most stereoscopic 3D (S3D) image visual discomfort predictors use the Support Vector Regressor (SVR) as the regression model. However, there are other good regression models such as the Random Forests (RF) and Gradient Boost Regression Tree (GBRT). Here we study the efficacy of these regression models for S3D image visual discomfort prediction. We deployed several regression models to predict the visual discomfort scores on S3D images using three kinds of features extracted from the images in the IEEE-SA and EPFL S3D image databases. So the prediction performance was evaluated to compare the performance of the different regression models. We also studied the issue of over-fitting which can affect model performance. The comparison results show that it is feasible and reliable to apply Random Forests or GBRT for S3D visual discomfort prediction with better performance than SVR.

关键词： Visualization Feature extraction Kernel Predictive models Vegetation Radio frequency Analytical models

来源：评论

学校读者我要写书评

暂无评论

Angel Girl of Visually Impaired Artists: Painting Navigation System for Blind or Visually Impaired Painters

Angel Girl of Visually Impaired Artists: Painting Navigation...

引用

IEEE Visual communications and image processing (VCIP)

作者： Hang Liu Menghan Hu Guangtao Zhai Huijing Huang Wa Zhang Qingli Li Yinghong Tian Yanling Shi Shanghai Key Laboratory of Multidimensional Information Processing East China Normal University China Institute of Image Communication and Information Processing Shanghai Jiao Tong University China

For those who love painting but unfortunately have visual impairments, holding a paintbrush to create a work is really a difficult task. For the purpose of solving this problem, a painting navigation system for visually impaired painters is introduced through the live demonstration. When painting, the developed system can endow visually impaired persons with the ability to perceive the surrounding environment, thus helping them realize their dream of painting. To achieve this goal, we designed four main modules viz., QR code based drawing board positioning module, brush real-time positioning module, color recognition module and human-computer interaction module, and integrated them into the system. In the validation experiments, the blindfolded users can successfully create a painting with the help of the developed navigation system. Moreover, the users told us that this system provided them with good experience. In a way, this painting navigation system can be seen as "angel's eyes" of visually impaired painters. The demo video of the proposed painting navigation system is available at: https://***/10.6084/***.9760004.v1.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An improved full-reference image quality metric based on structure compensation

An improved full-reference image quality metric based on str...

引用

Asia-Pacific Signal and Information processing Association Annual Summit and Conference (APSIPA)

作者： Ke Gu Guangtao Zhai Xiaokang Yang Wenjun Zhang Institute of Image Communication and Information Processing Shanghai Jiao Tong University Shanghai China Shanghai Key Laboratory of Digital Media Processingand Transmissions

During the last two decades, image quality assessment has been a major research area, which considerably helps to promote the development of image processing. Following the tremendous success of Structural SIMilarity (SSIM) index in terms of the correlation between the quality predictions and the subjective scores, many improved algorithms have been further exploited, such as Multi-Scale SSIM (MS-SSIM) and Information content Weighted SSIM (IW-SSIM). However, a growing number of researchers have been devoted to the study of the effects of uneven responses to different image distortion categories on prediction accuracy of the quality metrics. Inspired by this, we propose an improved full-reference image quality assessment paradigm based on structure compensation. Experimental results on laboratory for image and Video Engineering (LIVE) database and Tampere image Database 2008 (TID2008) are provided to confirm our introduced approach has superior prediction performance as compared to mainstream image quality metrics. Besides, it is worth emphasizing that our algorithm not introduces other operators but only applies the SSIM function to compensate itself, and furthermore, it also has an effective capability of image distortion classification.

关键词： Nonlinear distortion Measurement image quality Transform coding Indexes Prediction algorithms

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Embedded Multimedia Surveillance System

Design and Implementation of Embedded Multimedia Surveillanc...

引用

International Workshop on Knowledge Discovery and Data Mining (WKDD)

作者： Yan Liu Ren-Fa Li Cheng Xu Fei Yu School of Computer and Communication Hunan University Changsha China Guangdong Province Key Laboratory of Electronic Commerce Market Application Technology Guangdong University of Business Studies Guangzhou China Provincial Key Laboratory of Image Processing & Image Communication Nanjing University of Posts and Telecommunications Nanjing China

Due to the thrive of networking multimedia, artificial intelligence and embedded system, video surveillance system is evolved from computer-based to embedded-based. This paper presents an intelligent multimedia surveillance system based on embedded video server. Design and implementation issues of this surveillance system are discussed including system architecture, hardware design and software framework. Moreover, practical applications show that our framework can suitable for commercial applications.

关键词： Multimedia systems Artificial intelligence Application software Embedded system Video surveillance Computer networks Embedded computing Intelligent systems Network servers Computer architecture

来源：评论

学校读者我要写书评

暂无评论

Multilevel DCT-based zerotree image coding

Multilevel DCT-based zerotree image coding

引用

IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB)

作者： Kankan Zheng Liang Zhang Rong Xie Institute of Image Communication and Network Engineering & Shanghai Key Laboratory of Digital Media Processing and Transmission Shanghai Jiao Tong University Shanghai China

image and video coding is fundamental to data storage and transmission. JPEG 2000 is the latest international image coding standard which exploits wavelet transform for image compression. Zerotree coding structure is an efficient algorithm for coding of wavelet transform coefficients. However, implementation complexity of wavelet transform and long coding process of zerotree structure may limit the application in some cases. In this paper, an idea of multilevel DCT-based zerotree image coding is proposed, in which Wavelet Transform (WT) is replaced by Discrete Cosine Transform (DCT) and DCT coefficients are regrouped into subbands of different frequencies. Furthermore, a new zerotree coding structure is proposed to reduce the complexity of zerotree coding process. A conventional zerotree with a parent-child relation of 2×2 offspring is expanded to the one of 4×4 offspring. Coefficients reorder and inverse DCT are needed when DCT transform size is not a power of 4. This expansion makes each parent node of zerotree coding structure have more child nodes and, therefore, reduce the complexity of encoding process. Simulation results of images with different spatial resolutions are presented to show the advantage of the proposed algorithm. The proposed multi-level DCT-based zerotree image coding algorithm performs similar coding performance as the SPIHT method. The zerotree coding structure with a parent-child relation of 4×4 pixels can significantly reduce the complexity of high-bitrate coding process without sacrificing image subjective quality.

关键词： Discrete cosine transforms image coding Encoding Spatial resolution Wavelet transforms Bit rate

来源：评论

学校读者我要写书评

暂无评论

Online video synopsis method through simple tube projection strategy

Online video synopsis method through simple tube projection ...

引用

International Conference on Wireless communications and Signal processing (WCSP)

作者： Jing Jin Feng Liu Zongliang Gan Ziguan Cui Jiangsu Province Key Lab on Image Processing & Image Communications Nanjing University of Posts and Telecommunications Nanjing China Key Lab of Broadband Wireless Communication and Sensor Network Technology Ministry of Education Nanjing University of Posts and Telecommunications Nanjing China

ISBN: (纸本)9781509028610

Traditional video synopsis methods model the processing into an optimization formula where relations among objects such as collision cost are utilized while entire re-calculation is introduced under each possible temporal shift. Unlike the pairwise cost optimization, we propose a low-complexity and efficient online synopsis method where each tube is processed independently. Without tubes' comparison, the rearrangement is accomplished by a simple projection strategy and an updating projection matrix which records the newest information of the moving space. Furthermore, buffer and a predefined fitness condition also help to increase spatial and temporal utilization. Experiment results demonstrate that the proposed method is superior to other synopsis methods in the processing speed and temporal consistency.

关键词： Streaming media Electron tubes Trajectory Optimization Surveillance Telecommunications Cameras

来源：评论

学校读者我要写书评

暂无评论

Stereoscopic images quality assessment based on deep learning

Stereoscopic images quality assessment based on deep learnin...

引用

IEEE Visual communications and image processing (VCIP)

作者： Kai Wang Jun Zhou Ning Liu Xiao Gu Shanghai Key Laboratory of Digital Media Processing and Transmissions Shanghai China Institute of Image Communication and Network Engineering Shanghai Jiao Tong University

ISBN: (纸本)9781509053179

With the popularity of stereoscopic 3D (S3D) images and videos, many advanced objective quality assessment methods have been proposed to evaluate viewers' Quality of Experience (QoE). Among them, most algorithms take advantages of the disparity maps to extract useful features. On the other hand, deep learning has been one of the hottest research topics during these years, but limited efforts focused on the field in objective quality evaluation of S3D images. In this paper, we propose a S3D image quality assessment (S3D IQA) method based on deep learning. In this method, the Convolutional Restricted Boltzmann Machines (CRBM) combined with Factored Third-Order RBM (FTO-RBM) is considered as learning model to extract feature maps from pre-processed left and right images automatically. Then an improved traversal algorithm based on two pooling strategies is put forward to optimize extracted feature maps, which improves the final quality assessment performance significantly. Experimental results show that our S3D IQA method achieves good performance on 3D databases tested.

关键词： Feature extraction Training Databases Machine learning image quality Three-dimensional displays Data preprocessing

来源：评论

学校读者我要写书评

暂无评论

Spatial-Temporal Constrained Pseudo-labeling for Unsupervised Person Re-identification via GCN Inference 18th

Spatial-Temporal Constrained Pseudo-labeling for Unsupervis...

引用

18th International Forum of Digital Multimedia communication, IFTC 2021

作者： Ling, Sen Yang, Hua Liu, Chuang Chen, Lin Zhao, Hongtian The Institute of Image Communication and Network Engineering Department of Electronic Engineering Shanghai Jiao Tong University Shanghai China Shanghai Key Laboratory of Digital Media Processing and Transmission Shanghai Jiao Tong University Shanghai China MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai China

ISBN: (纸本)9789811922657

Most existing unsupervised person re-identification (Re-ID) methods primarily depend on the cluster distance, and merely exploit the available source labeled data to assign pseudo labels for the unannotated data. Whereas, the cluster distance usually fails to adapt to different datasets due to the domain gap. Besides, learning exclusively from the source data can not generate accurate pseudo labels for the lack of the target data information. To address this problem, we propose to exploit the spatial-temporal constraints to facilitate the pseudo label generation process. Specifically, graphs for the labeled source data are constructed and the graph convolution network (GCN) is used to learn graph embeddings. Based on these graph embeddings, the likelihood of linkages between graph nodes is estimated and utilized to assign pseudo labels for the unlabeled data. Then, with the pseudo labels, a smoothed spatial-temporal probability distribution model is generated to amend the likelihood of linkages between graph nodes as well as correct the visual similarity scores for person Re-ID. Finally, we optimize the pseudo label assignment, feature extraction networks, and spatial-temporal model alternatively and iteratively to improve the person Re-ID performance. Comprehensive experiments demonstrate that the proposed method outperforms state-of-the-art methods. © 2022, Springer Nature Singapore Pte Ltd.

关键词： Graph theory

来源：评论

学校读者我要写书评

暂无评论

Enhancement of color image based on tone-preserving

Enhancement of color image based on tone-preserving

引用

International Conference on Wireless communications and Signal processing (WCSP)

作者： Ya'nan Yang Xiaofan Wang Feng Liu Zongliang Gan Jiangsu Province Key Lab on Image Processing & Image Communications Nanjing University of Posts and Telecommunications Nanjing China Key Lab of Broadband Wireless Communication and Sensor Network Technology Ministry of Education Nanjing University of Posts and Telecommunications Nanjing China

ISBN: (纸本)9781509028610

Since the lighting conditions in strong contrast regions between the light and dark cant be estimated accurately by traditional center/surround Retinex algorithm, the over-enhancement and color distortion may exist. In view of this, combining with the human visual characteristics, a color image enhancement algorithm based on tone-preserving was proposed. A determination function was added to the bilateral filter to estimate illuminance image more accurately and weaken over-enhancement. According to human visual masking effect, the improved gamma correction was utilized to correct the brightness of illumination image adaptively and the local contrast of reflection image obtained by division was enhanced based on local statistics. Besides, the final enhanced image was obtained by combining illumination image with reflection image, which can make image appear more natural. Compared with other similar algorithms from both subjective and objective aspects, the results show that this method being applied to low-contrast color image enhancement can not only improve image clarity, but reduce color distortion.

关键词： image color analysis Brightness Lighting Color Visualization Nonlinear distortion

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：