检索结果-内蒙古大学图书馆

IEEE International conference on image Processing

作者： Hyeongmin Lee Taeoh Kim Eungyeol Song Sangyoun Lee School of Electrical and Electronic Engineering Yonsei University Republic of Korea

Designing models for learning dataset with complex distributions is one of the main challenges that still remains in machine learning areas. We propose CollaboNet, which can divide a large dataset into sub-datasets, train two generative models separately, and let two models work together to achieve better performance. The proposed algorithm divides a large dataset without label since the capability difference between two generative models in performing tasks on each data is the main criterion for dividing a large dataset. In other words, the classification model can be trained by unsupervised manner. Autoencoder experiments for pure MNIST and the datasets combined artificially from two image sets shows that CollaboNet successfully splits large datasets without labels, improving the performance of generative models.

关键词： Task analysis Mathematical model data models Manganese image reconstruction Cats Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

DualGAN: Unsupervised Dual Learning for image-to-image Translation 16

DualGAN: Unsupervised Dual Learning for Image-to-Image Trans...

引用

16th IEEE International conference on Computer vision (ICCv)

作者： Yi, Zili Zhang, Hao Tan, Ping Gong, Minglun Mem Univ Newfoundland St John NF Canada Simon Fraser Univ Burnaby BC Canada

ISBN: (纸本)9781538610329

Conditional Generative Adversarial Networks( GANs) for cross-domain image-to-image translation have made much progress recently [7, 8, 21, 12, 4, 18]. Depending on the task complexity, thousands to millions of labeled image pairs are needed to train a conditional GAN. However, human labeling is expensive, even impractical, and large quantities of data may not always be available. Inspired by dual learning from natural language translation [23], we develop a novel dual-GAN mechanism, which enables image translators to be trained from two sets of unlabeled images from two domains. In our architecture, the primal GAN learns to translate images from domain U to those in domain v, while the dual GAN learns to invert the task. The closed loop made by the primal and dual tasks allows images from either domain to be translated and then reconstructed. Hence a loss function that accounts for the reconstruction error of images can be used to train the translators. Experiments on multiple image translation tasks with unlabeled data show considerable performance gain of DualGAN over a single GAN. For some tasks, DualGAN can even achieve comparable or slightly better results than conditional GAN trained on fully labeled data.

关键词： Gallium nitride Generators image reconstruction Training data Training Computer vision

来源：评论

学校读者我要写书评

暂无评论

SYNTHESIZING PHOTO-REALISTIC 3D TALKING HEAD: LEARNING LIP SYNCHRONICITY AND EMOTION from AUDIO AND vIDEO

SYNTHESIZING PHOTO-REALISTIC 3D TALKING HEAD: LEARNING LIP S...

引用

IEEE International conference on image Processing

作者： Jun Yu Lingyun Yu Department of Automation University of Science and Technology of China

We propose a speech driven emotional 3D talking head. First, based on face prior knowledge, an individual head model is reconstructed by integrating the internal articulators from Magnetic Resonance Imaging slices with the appearance from visible images. Second, the blendshapes of each phoneme and the facial expression are synthesized by statistical learning. Third, visual coarticulation is modeled by training a deep neural network on parallel audio-visual data. Finally, the articulatory animations of continuous phonemes are fused by visual coarticulation model and facial expressions to produce expressive speech synchronized animations. Experiments show that the system can not only distinguish the visual differences among phonemes in real-time, but also significantly increase the ability of human perception.

关键词： Animation Hidden Markov models Head image reconstruction Three-dimensional displays visualization Skin

来源：评论

学校读者我要写书评

暂无评论

CARISI: CONvOLUTIONAL AUTOENCODER-BASED INTER-SLICE INTERPOLATION OF BRAIN TUMOR vOLUMETRIC imageS

CARISI: CONVOLUTIONAL AUTOENCODER-BASED INTER-SLICE INTERPOL...

引用

IEEE International conference on image Processing

作者： Parnian Afshar Atefeh Shahroudnejad Arash Mohammadi Konstantinos N. Plataniotis Concordia Institute for Information Systems Engineering Concordia University Montreal QC Canada Department of Electrical and Computer Engineering University of Toronto Toronto ON Canada

The paper is motivated by the fact that brain cancer is one of the deadliest cancers and its detection in early stages is of paramount importance. In this regard, tumor 3D shape reconstruction from magnetic resonance (MR) or computed tomography (CT) scans provides critical information, which can not be interpreted from 2D images. However, CT and MR images have low resolution in z direction compared to their resolution in x and y directions, therefore, 3D reconstructed shapes are of low quality. In this paper, we propose to use convolutional auto-encoders (CAEs) to address this drawback, and develop a convolutional autoencoder-based inter-slice interpolation (CARISI) framework. Although deep nets have been used very recently for brain tumor segmentation, to the best of our knowledge, this is the first attempt to use CAEs for 3D reconstruction of brain tumor. The proposed CARISI framework consists of several encoding and decoding components, which can handle rapid changes in tumor shape without the need for supervision of an expert. Our experiments based on a real data-set consisting of 3064 segmented brain tumor images indicate that the proposed CARISI framework outperforms its counterpart and has the potential to significantly improve the overall quality of the reconstructed shapes.

关键词： Tumors Three-dimensional displays Interpolation image reconstruction Shape Computed tomography Convolution

来源：评论

学校读者我要写书评

暂无评论

An improved compressed sensing method for cDNA miacorarray image denoising

An improved compressed sensing method for cDNA miacorarray i...

引用

Chinese Control conference

作者： Guifang Shao Ao Zhang Weiting Zhang Junmei Zou Yiran Wang Shengjia Zheng Department of Automation Xiamen University Xiamen China School of Information Engineering Jimei University Xiamen China

Microarray image processing plays an important role among array analysis and its precision will affect the application of gene chip. However, real microarray images drawn from various experimental procedures may contain noises and other artifacts, which result in a lower image processing accuracy. Hence, in this paper, we propose a novel image denoising method by introducing the compressed sensing theory into microarray image noise reduction. To improve the matching pursuit efficiency, the correlation coefficient is adopted. Experiments on 17 simulated images and 59 real microarray images obtained from the DeRisi, SIB and BCM data sets are conducted. Qualitative and quantitative results compared with the Donoho threshold, generalized wavelet and adaptive thresholding algorithm, reveal that our proposed method performs better on reducing the noises.

关键词： Noise reduction Matching pursuit algorithms Correlation coefficient Sparse matrices image reconstruction image denoising

来源：评论

学校读者我要写书评

暂无评论

image Generation and Translation with Disentangled Representations

Image Generation and Translation with Disentangled Represent...

引用

International Joint conference on Neural Networks

作者： Tobias Hinz Stefan Wermter Department of Informatics Universität Hamburg Hamburg Germany

Generative models have made significant progress in the tasks of modeling complex data distributions such as natural images. The introduction of Generative Adversarial Networks (GANs) and auto-encoders lead to the possibility of training on big data sets in an unsupervised manner. However, for many generative models it is not possible to specify what kind of image should be generated and it is not possible to translate existing images into new images of similar domains. Furthermore, models that can perform image-to-image translation often need distinct models for each domain, making it hard to scale these systems to multiple domain image-to-image translation. We introduce a model that can do both, controllable image generation and image-to-image translation between multiple domains. We split our image representation into two parts encoding unstructured and structured information respectively. The latter is designed in a disentangled manner, so that different parts encode different image characteristics. We train an encoder to encode images into these representations and use a small amount of labeled data to specify what kind of information should be encoded in the disentangled part. A generator is trained to generate images from these representations using the characteristics provided by the disentangled part of the representation. Through this we can control what kind of images the generator generates, translate images between different domains, and even learn unknown data-generating factors while only using one single model.

关键词： Generators Training image generation Gallium nitride data models Face image reconstruction

来源：评论

学校读者我要写书评

暂无评论

data-DRIvEN SYNTHETIC CEREBROvASCULAR MODELS FOR vALIDATION OF SEGMENTATION ALGORITHMS

DATA-DRIVEN SYNTHETIC CEREBROVASCULAR MODELS FOR VALIDATION ...

引用

Annual International conference of the IEEE Engineering in Medicine and Biology Society

作者： Michael R. Nowak Yoonsuck Choe Texas A&M University College Station College Station TX US

We introduce a novel method to generate biologically grounded synthetic cerebrovasculature models in a data-driven fashion. First, the centerlines of vascular filaments embedded in an acquired imaging volume are obtained by a segmentation algorithm. That imaging volume is reconstructed from a graph encoding of the centerline (i.e., generating the model's ground truth) and the segmentation algorithm is applied to the resultant volume. As the location and characteristics of the vasculature embedded in this volume are known, the accuracy of the segmentation algorithm can be assessed. Moreover, because the synthetic volume was reconstructed directly from biological data, an assessment is made on embedded filaments that are representative of the topological and geometrical characteristics of the dataset. We believe that such models will provide the means necessary for the enhanced evaluation of vascular segmentation algorithms.

关键词： Biological system modeling image segmentation Biological information theory Imaging Topology Manuals image reconstruction

来源：评论

学校读者我要写书评

暂无评论

A fast gradient projection method for 3D image reconstruction from limited tomographic data 7

A fast gradient projection method for 3D image reconstructio...

引用

7th International conference on New Computational Methods for Inverse Problems (NCMIP)

作者： Coli, v. L. Piccolomini, E. Loli Morotti, E. Zanni, L. Univ Modena & Reggio Emilia Dept Phys Informat & Math Modena Italy Univ Bologna Dept Math Bologna Italy Univ Padua Dept Math Padua Italy

We consider in this paper the problem of reconstructing 3D Computed Tomography images from limited data. The problem is modeled as a nonnegatively constrained minimization problem of very large size. In order to obtain an acceptable image in short time, we propose a scaled gradient projection method, accelerated by exploiting a suitable scaling matrix and efficient rules for the choice of the step-length. In particular, we select the step-length either by alternating Barzilai-Borwein rules or by exploiting a limited number of back gradients for approximating second-order information. Numerical results on a 3D Shepp-Logan phantom are presented and discussed.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

DISTRIBUTED ONE-CLASS LEARNING

DISTRIBUTED ONE-CLASS LEARNING

引用

IEEE International conference on image Processing

作者： Ali Shahin Shamsabadi Hamed Haddadi Andrea Cavallaro Queen Mary University of London Llrnperial College London

We propose a cloud-based filter trained to block third parties from uploading privacy-sensitive images of others to online social media. The proposed filter uses Distributed One-Class Learning, which decomposes the cloud-based filter into multiple one-class classifiers. Each one-class classifier captures the properties of a class of privacy-sensitive images with an autoencoder. The multi-class filter is then reconstructed by combining the parameters of the one-class autoencoders. The training takes place on edge devices (e.g. smartphones) and therefore users do not need to upload their private and/or sensitive images to the cloud. A major advantage of the proposed filter over existing distributed learning approaches is that users cannot access, even indirectly, the parameters of other users. Moreover, the filter can cope with the imbalanced and complex distribution of the image content and the independent probability of addition of new users. We evaluate the performance of the proposed distributed filter using the exemplar task of blocking a user from sharing privacy-sensitive images of other users. In particular, we validate the behavior of the proposed multi-class filter with non-privacy-sensitive images, the accuracy when the number of classes increases, and the robustness to attacks when an adversary user has access to privacy-sensitive images of other users.

关键词： Training image reconstruction Privacy Training data Feature extraction image edge detection data privacy

来源：评论

学校读者我要写书评

暂无评论

Single image super resolution based on sparse domain selection

引用

NEUROCOMPUTING 2017年 269卷 180-187页

作者： Lu, Wen Sun, Huxing Wang, Rui He, Lihuo Jou, MingJong Syu, ShenSian Li, JiShiang Xidian Univ Sch Elect Engn Xian 710071 Shaanxi Peoples R China Shenzhen China Star Optoelect Technol Co Ltd Shenzhen Guangdong Peoples R China

Single image super-resolution (SR) aims to form a high-resolution (HR) image from an input low resolution (LR) image. The sparse coding-based example learning methods typically assume that the low-resolution and high-resolution features share the same representation coefficients over their own dictionaries. The assumption is a strong constraint which limits the flexibility to model the complexity of feature space and the mapping among them. To solve the problem, this paper proposes a novel single image super-resolution method utilizing sparse domain selection. In the training phase, the efficient mapping between LR and HR coefficients is established by searching the sparse domain among feature spaces spanned by LR-HR dictionaries. Then this mapping and learning HR dictionary are optimized jointly through minimizing the sparse representation error and sparse domain mapping error. During the reconstruction phase, the learned mapping from the input LR feature is applied to the desired HR feature to achieve accurate and stable SR recovery. Experimental results indicate that the proposed approach is more capable of modeling the relationships from feature of LR to those of HR, thus improves the quality of reconstruction image. (C) 2017 Elsevier B.v. All rights reserved.

关键词： Single image super-resolution Example learning-based Sparse domain selection

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：