基于DCT变换和零次学习的刑侦图像超分辨率

图像超分辨率在视频侦查领域有重要作用.基于卷积神经网络的超分辨率算法通常在训练时输入人工合成的低分辨率图像,学习高、低分辨率图像的映射,很难应用于视频侦查领域.真实低分辨率图像退化过程复杂未知,且大都经过压缩算法的处理,存在人工压缩痕迹,导致超分辨率图像出现假纹理.针对真实场景下的低分辨率图像提出一种基于离散余弦变换(...     

基于自适应低秩去噪的磁共振图像重构

本文提出了一种基于自适应低秩去噪的磁共振图像重构算法.该方法使用去噪近似消息传递算法重构磁共振图像,将自适应加权Schatten-p范数最小化方法 (Weighted Schatten p-Norm Minimization, WSNM)作为其降噪模型,研究图像的重构性能.根据算法迭代过程中估计的噪声标准差自适应的设定WSNM的图像块大小及相似块个数.实验表明,与近几年提出的磁共振图像重构算法比较,本文提出的算法可以获得更高的峰值信噪比(Peak Signal to Noise Ratio, PSNR)和更低的相对L₂范数误差(Relative L₂ Norm Error, RLNE),得到更好的重建效果.     

A reinforcement agent for object segmentation in ultrasound images

The principal contribution of this work is to design a general framework for an intelligent system to extract one object of interest from ultrasound images. This system is based on reinforcement learning. The input image is divided into several sub-images, and the proposed system finds the appropriate local values for each of them so that it can extract the object of interest. The agent uses some images and their ground-truth (manually segmented) version to learn from. A reward function is employed to measure the similarities between the output and the manually segmented images, and to provide feedback to the agent. The information obtained can be used as valuable knowledge stored in the Q-matrix. The agent can then use this knowledge for new input images. The experimental results for prostate segmentation in trans-rectal ultrasound images show high potential of this approach in the field of ultrasound image segmentation.     

基于图像局部奇异值向量和BP神经网络分类器的道路导航方法

提出以道路图像矩阵的局部奇异值向量作为特征输入,以BP神经网络作为分类器的道路导航方法.首先将图像分割成若干子图像,然后分别对子图像进行奇异值分解,提取子图像的代数特征向量.子图像的特征奇异值组成整个图像的局部奇异值向量,作为分类器的输入.再利用BP神经网络分类器对道路图像进行训练及识别.实验中处理了三类道路图像(偏左、偏右、正确方向),每类用20 幅图像作为训练样本,30幅用作测试.结果表明,这 种道路导航方法的识别率达到了100%.􀁳 􀁨 􀁱     

Edge preserving mixed noise removal

To faithfully recover the clean images corrupted by additive white Gaussian noise (AWGN) and impulse noise (IN), a novel edge preserving image denoising algorithm is proposed. The low- and high-frequency components of the image are restored separately. The high-frequency components of the images are restored based on nonlocal self-similarity (NSS) learning from natural images. An energy minimization function is developed to combine the low- and high-frequency components into one model. Experiments demonstrate that the proposed method outperforms existing mixture noise removal methods in peak signal-to-noise ratio (PSNR), edges preservation and visual performance.

     

基于NSST域卷积神经网络的低剂量CT图像恢复

为解决低剂量CT（Low-Dose Computed Tomography,LDCT）图像中的噪声/伪影问题,提出一种基于非下采样Shearlet变换（Non-Sample Shearlet Transformation,NSST）的卷积神经网络（Convolution Neural Network,CNN）的NSST-CNN模型。训练时,对数据集中的常规剂量CT（Normal-Dose Computed Tomography,NDCT）和LDCT图像做NSST分解,将LDCT图像的高频子图作为输入,LDCT和NDCT图像的高频子图的残差图像作为标签,通过CNN训练,学习LDCT高频子图和高频残差子图的映射关系;测试时,将LDCT图像的高频子图减去利用映射关系预测的主要包括噪声/伪影的高频子图,然后做NSST反变换得到高质量的LDCT图像。实验结果表明,与KSVD、BM3D以及图像域CNN方法相比,NSST-CNN模型得到的结果具有更高的峰值信噪比和结构相似度,更接近NDCT图像。     

沙尘图像色彩恢复及增强卷积神经网络

目的 在沙尘天气条件下,由于大气中悬浮微粒对入射光线的吸收和散射,户外计算机视觉系统所采集图像通常存在颜色偏黄失真和低对比度等问题,严重影响户外计算机视觉系统的性能。为此,提出一种带色彩恢复的沙尘图像卷积神经网络增强方法,由一个色彩恢复子网和一个去尘增强子网组成。方法 采用提出的色彩恢复子网(sand dust color correction, SDCC)校正沙尘图像的偏色,将颜色校正后的图像作为条件,输入到由自适应实例归一化残差块组成的去尘增强子网中,对沙尘图像进行增强处理。本文还提出一种基于物理光学模型的沙尘图像合成方法,并采用该方法构建了大规模的配对沙尘图像数据集。结果 对大量沙尘图像的实验结果表明,所提出的沙尘图像增强方法能很好地去除图像中的偏色和沙尘,获得正常的视觉颜色和细节清晰的图像。进一步的对比实验表明,该方法能取得优于对比方法的增强图像。结论 本文所提出的沙尘图像增强方法能很好地消除整体的黄色色调和尘霾现象,获得正常的视觉色彩和细节清晰的图像。     

小波变换与纹理合成相结合的图像修复

目的 为了克服传统的图像修复算法在结构和纹理边界的错误修复,利用小波变换域的系数特征,探讨了一种基于小波变换与纹理合成相结合的修复算法。方法 算法先利用小波变换将待修复图像分解成具有不同分辨率的低频子图和高频子图,然后根据不同子图各自的特征分别进行修复。对代表图像结构信息的低频子图,采用FMM(fast marching method)算法进行修复;对代表图像纹理信息的高频子图,根据各子图中小波系数的特征,利用纹理合成方法进行修复。结果 分层、分类修复方法对边缘破损具有良好的修复效果,其峰值信噪比相比于传统算法提高了1~2 dB。结论 与相关算法相比,本文算法的综合修复能力较好,可以有效修复具有较强边缘和丰富纹理的破损图像,尤其对破损自然图像的修复,修复后图像质量得到较大提升,修复效果更符合人眼视觉效应。     

结合内外先验知识的低照度图像增强与去噪算法

目的 现有大多数低照度图像增强算法会放大噪声,且用于极低照度图像时会出现亮度提升不足、色彩失真等问题。为此,提出一种基于Retinex(retina cortex)的增强与去噪方法。方法 为了增强极低照度图像,首先利用暗通道先验原理估计场景的全局光照,若光照低于0.5,对图像进行初始光照校正;其次,提出一种Retinex顺序分解模型,使低照度图像中的噪声均体现在反射分量中,基于分解结果,利用Gamma校正求取增强后的噪声图像;最后,提出一种基于内外双重互补先验约束的去噪机制,利用非局部自相似性原理为反射分量构建内部先验约束,基于深度学习,为增强后的噪声图像构建外部先验约束,使内外约束相互制约。结果 将本文算法与6种算法比较,在140幅普通低照度图像和162幅极低照度图像上（有正常曝光参考图像）进行主观视觉和客观指标评价比较,结果显示本文方法在亮度提升、色彩保真及去噪方面均有明显优势,对于普通低照度图像,BTMQI(blind tone-mapped quality index)和NIQE(natural image quality evaluator)指标均取得次优值,对于极低照度图像...     

残差密集相对平均CGAN的脑部图像配准

目的 针对图像合成配准算法中鲁棒性差及合成图像特征信息不足导致配准精度不高的问题,提出了基于残差密集相对平均条件生成对抗网络（residual dense-relativistic average conditional generative adversarial network,RD-RaCGAN）的多模态脑部图像配准方法。方法 相对平均生成对抗网络中的相对平均鉴别器能够增强模型稳定性,条件生成对抗网络加入条件变量能提高生成数据质量,结合两种网络特点,利用残差密集块充分提取深层网络特征的能力,构建RD-RaCGAN合成模型。然后,待配准的参考CT（computed tomography）和浮动MR（magnetic resonance）图像通过已训练好的RD-RaCGAN合成模型双向合成对应的参考MR和浮动CT图像。采用区域自适应配准算法,从参考CT和浮动CT图像中选取骨骼信息的关键点,从浮动MR和参考MR图像中选取软组织信息的关键点,通过提取的关键点指导形变场的估计。从浮动CT图像到参考CT图像估计一个形变场。类似地,从浮动MR图像到参考MR图像估计一个形变场。另外,采用分层对称的思想进一步优化两个形变场,当两个形变场之间的差异达到最小时,将两个形变场融合得到最终的形变场,并将形变场作用于浮动图像完成配准。结果 实验结果表明,与其他6种图像合成方法相比,本文模型合成的目标图像在视觉效果和客观评价指标上均优于其他方法。对比Powell优化的MI（mutual information）法、ANTs-SyN（advanced normalization toolbox-symmetric normalization）、D.Demons（diffeomorphic demons）、Cue-Aware Net（cue-aware deep regression network）和I-SI（intensity and spatial information）的图像配准方法,归一化互信息分别提高了43.71%、12.87%、10.59%、0.47%、5.59%,均方根误差均值分别下降了39.80%、38.67%、15.68%、4.38%、2.61%。结论 本文提出的多模态脑部图像配准方法具有很强的鲁棒性,能够稳定、准确地完成图像配准任务。     

残差密集注意力网络多模态MR图像超分辨率重建

目的 现有医学图像超分辨率方法主要针对单一模态图像进行设计,然而在磁共振成像(magnetic resonance imaging, MRI)技术的诸多应用场合,往往需要采集不同成像参数下的多模态图像。针对单一模态的方法无法利用不同模态图像之间的关联信息,很大程度上限制了重建性能。目前超分辨率网络模型参数量往往较大,导致计算和存储代价较高。为此,本文提出了一个轻量级残差密集注意力网络,以一个统一的网络模型同时实现多模态MR图像的超分辨率重建。方法 首先将不同模态的MR图像堆叠后输入网络,在低分辨率空间中提取共有特征,之后采用设计的残差密集注意力模块进一步精炼特征,再通过一个亚像素卷积层上采样到高分辨率空间,最终分别重建出不同模态的高分辨率图像。结果 本文采用MICCAI (medical image computing and computer assisted intervention) BraTS (brain tumor segmentation) 2019数据集中的T1和T2加权MR图像对网络进行训练和测试,并与8种代表性超分辨率方法进行对比。实验结果表明,本文方法可以取得优于...     

Guided Hyperspectral Image Denoising with Realistic Data

The hyperspectral image (HSI) denoising has been widely utilized to improve HSI qualities. Recently, learning-based HSI denoising methods have shown their effectiveness, but most of them are based on synthetic dataset and lack the generalization capability on real testing HSI. Moreover, there is still no public paired real HSI denoising dataset to learn HSI denoising network and quantitatively evaluate HSI methods. In this paper, we mainly focus on how to produce realistic dataset for learning and evaluating HSI denoising network. On the one hand, we collect a paired real HSI denoising dataset, which consists of short-exposure noisy HSIs and the corresponding long-exposure clean HSIs. On the other hand, we propose an accurate HSI noise model which matches the distribution of real data well and can be employed to synthesize realistic dataset. On the basis of the noise model, we present an approach to calibrate the noise parameters of the given hyperspectral camera. Besides, on the basis of observation of high signal-to-noise ratio of mean image of all spectral bands, we propose a guided HSI denoising network with guided dynamic nonlocal attention, which calculates dynamic nonlocal correlation on the guidance information, i.e., mean image of spectral bands, and adaptively aggregates spatial nonlocal features for all spectral bands. The extensive experimental results show that a network learned with only synthetic data generated by our noise model performs as well as it is learned with paired real data, and our guided HSI denoising network outperforms state-of-the-art methods under both quantitative metrics and visual quality.

     

Visual Perception in Realistic Image Synthesis

Realism is often a primary goal in computer graphics imagery, and we strive to create images that are perceptually indistinguishable from an actual scene. Rendering systems can now closely approximate the physical distribution of light in an environment. However, physical accuracy does not guarantee that the displayed images will have authentic visual appearance. In recent years the emphasis in realistic image synthesis has begun to shift from the simulation of light in an environment to images that look as real as the physical environment they portray. In other words the computer image should be not only physically correct but also perceptually equivalent to the scene it represents. This implies aspects of the Human Visual System (HVS) must be considered if realism is required. Visual perception is employed in many different guises in graphics to achieve authenticity. Certain aspects of the visual system must be considered to identify the perceptual effects that a realistic rendering system must achieve in order to reproduce effectively a similar visual response to a real scene. This paper outlines the manner in which knowledge about visual perception is increasingly appearing in state‐of‐the‐art realistic image synthesis. After a brief overview of the HVS, this paper is organized into four sections, each exploring the use of perception in realistic image synthesis, each with slightly different emphasis and application. First, Tone Mapping Operators, which attempt to map the vast range of computed radiance values to the limited range of display values, are discussed. Then perception based image quality metrics, which aim to compare images on a perceptual rather than physical basis, are presented. These metrics can be used to evaluate, validate and compare imagery. Thirdly, perception driven rendering algorithms are described. These algorithms focus on embedding models of the HVS directly into global illumination computations in order to improve their efficiency. Finally, techniques for comparing computer graphics imagery against the real world scenes they represent are discussed.     

GPU implementation of non-local maximum likelihood estimation method for denoising magnetic resonance images

Magnetic resonance imaging (MRI) is a widely deployed medical imaging technique used for various applications such as neuroimaging, cardiovascular imaging and musculoskeletal imaging. However, MR images degrade in quality due to noise. The magnitude MRI data in the presence of noise generally follows a Rician distribution if acquired with single-coil systems. Several methods are proposed in the literature for denoising MR images corrupted with Rician noise. Amongst the methods proposed in literature for denoising MR images corrupted with Rician noise, the non-local maximum likelihood methods (NLML) and its variants are popular. In spite of the performance and denoising quality, NLML algorithm suffers from a tremendous time complexity \(O\left( {m^{3} N^{3} } \right)\), where \(m^{3}\) and \(N^{3}\) represent the search window and image size, respectively, for a 3D image. This makes the algorithm challenging for deployment in the real-time applications where fast and prompt results are required. A viable solution to this shortcoming would be the application of a data parallel processing framework such as Nvidia CUDA so as to utilize the mutually exclusive and computationally intensive calculations to our advantage. The GPU-based implementation of NLML-based image denoising achieves significant speedup compared to the serial implementation. This research paper describes the first successful attempt to implement a GPU-accelerated version of the NLML algorithm. The main focus of the research was on the parallelization and acceleration of one computationally intensive section of the algorithm so as to demonstrate the execution time improvement through the application of parallel processing concepts on a GPU. Our results suggest the possibility of practical deployment of NLML and its variants for MRI denoising.     

面向纹理平滑的方向性滤波尺度预测模型

目的 传统图像处理的纹理滤波方法难以区分梯度较强的纹理与物体的结构,而深度学习方法使用的训练集生成方式不够合理,且模型表示方式比较粗糙,为此本文设计了一种面向纹理平滑的方向性滤波尺度预测模型,并生成了含有标签的新的纹理滤波数据集。方法 在现有结构图像中逐连通区域填充多种纹理图,生成有利于模型训练的纹理滤波数据集。设计了方向性滤波尺度预测模型,该模型包含尺度感知子网络和图像平滑子网络。前者预测得到的滤波尺度图不但体现了该像素与周围像素是否为同一纹理,而且还隐含了该像素是否为结构像素的信息。后者以滤波尺度图和原图的堆叠作为输入,凭借少量的卷积层快速得出纹理滤波的结果。结果 在本文的纹理滤波数据集上与7个算法进行比较,峰值信噪比(peak signal to noise ratio, PSNR)与结构相似度(structural similarity, SSIM)分别高于第2名2.79 dB、0.0133,均方误差(mean squared error, MSE)低于第2名6.863 8,运算速度快于第2名0.002 s。在其他数据集上的实验对比也显示出本文算法更好地保持结构与平滑纹理。通过...     

多尺度判别条件生成对抗网络的前列腺MRI图像分割方法

目的 由MRI（magnetic resonance imaging）得到的影像具有分辨率高、软组织对比好等优点,使得医生能更精确地获得需要的信息,精确的前列腺MRI分割是计算机辅助检测和诊断算法的必要预处理阶段。因此临床上需要一种自动或半自动的前列腺分割算法,为各种各样的临床应用提供具有鲁棒性、高质量的结果。提出一种多尺度判别条件生成对抗网络对前列腺MRI图像进行自动分割以满足临床实践的需求。方法 提出的分割方法是基于条件生成对抗网络,由生成器和判别器两部分组成。生成器由类似U-Net的卷积神经网络组成,根据输入的MRI生成前列腺区域的掩膜;判别器是一个多尺度判别器,同一网络结构,输入图像尺寸不同的两个判别器。为了训练稳定,本文方法使用了特征匹配损失。在网络训练过程中使用对抗训练机制迭代地优化生成器和判别器,直至判别器和生成器同时收敛为止。训练好的生成器即可完成前列腺MRI分割。结果 实验数据来自PROMISE12前列腺分割比赛和安徽医科大学第一附属医院,以Dice相似性系数和Hausdorff距离作为评价指标,本文算法的Dice相似性系数为88.9%,Hausdorff距离为5.3 mm,与U-Net、DSCNN（deeply-supervised convolutional neured network）等方法相比,本文算法分割更准确,鲁棒性更高。在测试阶段,每幅图像仅需不到1 s的时间即可完成分割,超出了专门医生的分割速度。结论 提出了一种多尺度判别条件生成对抗网络来分割前列腺,从定量和定性分析可以看出本文算法的有效性,能够准确地对前列腺进行分割,达到了实时分割要求,符合临床诊断和治疗需求。     

基于颈动脉粥样硬化斑块的多序列MR图像去噪与配准

目的 颈动脉粥样硬化斑块成分识别是预测脑血管疾病及脑卒中临床病发的主要依据,然而在实际中斑块成分识别会受到磁共振图像固有的噪声和多序列图像空间坐标不匹配的影响,为了更加精确地识别斑块的组成成分,提出基于小波分析和中值滤波相结合的去噪方法,以及基于图形上下文(shape context)的颈动脉多序列磁共振(MR)图像配准算法。方法 针对MR图像去噪,首先在传统小波去噪的基础上改进了阈值函数,根据高、低频子带噪声分布比重自适应的选取阈值。多序列MR图像配准方面,在提取多序列图像颈动脉血管边缘的基础上,用shape context描述子做血管形状匹配,依据产生的匹配点对进行迭代校正,计算参考图像与浮动图像的血管形变场,然后采用样条插值方法得到最终的配准结果。结果 利用本文去噪方法能有效地去除图像高、低频域的噪声,同时保护图像的原始细节。本文多序列MR图像血管配准方法使得配准后血管重合度达到了96%±0.8%。本文方法能够有效地提高MR图像的质量,验证了算法的有效性。结论 本文方法能够有效去除多序列磁共振图像噪声和空间位置不匹配的情况,本文去噪方法也适用于其他模态(CT,超声等)医学图像的斑点噪声以及噪声分布不均匀等情况下的噪声去除,本文配准方法也能够有效地处理基于小目标,对精度要求较高的精细配准问题。     

Exact Histogram Specification for Digital Images Using a Variational Approach

We consider the problem of exact histogram specification for digital (quantized) images. The goal is to transform the input digital image into an output (also digital) image that follows a prescribed histogram. Classical histogram modification methods are designed for real-valued images where all pixels have different values, so exact histogram specification is straightforward. Digital images typically have numerous pixels which share the same value. If one imposes the prescribed histogram to a digital image, usually there are numerous ways of assigning the prescribed values to the quantized values of the image. Therefore, exact histogram specification for digital images is an ill-posed problem. In order to guarantee that any prescribed histogram will be satisfied exactly, all pixels of the input digital image must be rearranged in a strictly ordered way. Further, the obtained strict ordering must faithfully account for the specific features of the input digital image. Such a task can be realized if we are able to extract additional representative information (called auxiliary attributes) from the input digital image. This is a real challenge in exact histogram specification for digital images. We propose a new method that efficiently provides a strict and faithful ordering for all pixel values. It is based on a well designed variational approach. Noticing that the input digital image contains quantization noise, we minimize a specialized objective function whose solution is a real-valued image with slightly reduced quantization noise, which remains very close to the input digital image. We show that all the pixels of this real-valued image can be ordered in a strict way with a very high probability. Then transforming the latter image into another digital image satisfying a specified histogram is an easy task. Numerical results show that our method outperforms by far the existing competing methods.     

基于生成对抗网络的MR图像合成研究

目前,深度学习已广泛用于MR医学图像分析,然而获取大量的MR训练数据集是一项巨大的挑战。针对传统基于物理方法的数据集扩充方法(如旋转、缩放、仿射变换等)均存在较大的局限性,提出一种新的图像合成器。首先对MR图像数据集进行统一的尺寸处理,然后将每张图像分为大小相等的非重叠块,再利用自动编码器提取图像块的编码,并将图像块编码和随机噪声作为生成器的输入,最后将生成的图像块拼接成完整的MR图像。采用该方法对脑部MR图像进行实验,并对生成图像进行定性定量分析。实验结果表明,所提方法能生成高质量的MR图像且优于其他主流方法。