基于U-Net的盐体识别方法

卷积神经网络在计算机视觉领域取得重大突破，利用其强大的图像处理能力，将地下沉积盐体的识别问题转化为图像语义分割问题，应用深度卷积神经网络实现盐体地震图像的像素级语义分割.本文在U-Net基础上，增加网络深度并同时引入批归一化和Dropout处理，使得神经网络模型具有更高的可信度和更强的泛化能力.通过实验发现，在卷积层之后引入批归一化处理，并在池化层和叠加层之后引入Dropout可以稳定提升模型对盐体图像的分割性能.     

Nonlocal maximum likelihood estimation method for denoising multiple-coil magnetic resonance images

Effective denoising is vital for proper analysis and accurate quantitative measurements from magnetic resonance (MR) images. Even though many methods were proposed to denoise MR images, only few deal with the estimation of true signal from MR images acquired with phased-array coils. If the magnitude data from phased array coils are reconstructed as the root sum of squares, in the absence of noise correlations and subsampling, the data is assumed to follow a non central-χ distribution. However, when the k-space is subsampled to increase the acquisition speed (as in GRAPPA like methods), noise becomes spatially varying. In this note, we propose a method to denoise multiple-coil acquired MR images. Both the non central-χ distribution and the spatially varying nature of the noise is taken into account in the proposed method. Experiments were conducted on both simulated and real data sets to validate and to demonstrate the effectiveness of the proposed method.     

An anisotropic images segmentation and bias correction method

Intensity inhomogeneities cause considerable difficulty in the quantitative analysis of magnetic resonance (MR) images. Thus, bias field correction is a necessary step before quantitative analysis of MR data can be undertaken. This paper presents an anisotropic approach to bias correction and segmentation for images with intensity inhomogeneities and noise. Intensity-based methods are usually applied to estimate the bias field; however, most of them only concern the intensity information. When the images have noise or slender topological objects, these methods cannot obtain accurate results or bias fields. We use structure information to construct an anisotropic Gibbs field and combine the anisotropic Gibbs field with the Bayesian framework to segment images while estimating the bias fields. Our method is able to capture bias of quite general profiles. Moreover, it is robust to noise and slender topological objects. The proposed method has been used for images of various modalities with promising results.     

DeepcomplexMRI: Exploiting deep residual network for fast parallel MR imaging with complex convolution

This paper proposes a multi-channel image reconstruction method, named DeepcomplexMRI, to accelerate parallel MR imaging with residual complex convolutional neural network. Different from most existing works which rely on the utilization of the coil sensitivities or prior information of predefined transforms, DeepcomplexMRI takes advantage of the availability of a large number of existing multi-channel groudtruth images and uses them as target data to train the deep residual convolutional neural network offline. In particular, a complex convolutional network is proposed to take into account the correlation between the real and imaginary parts of MR images. In addition, the k-space data consistency is further enforced repeatedly in between layers of the network. The evaluations on in vivo datasets show that the proposed method has the capability to recover the desired multi-channel images. Its comparison with state-of-the-art methods also demonstrates that the proposed method can reconstruct the desired MR images more accurately.     

A 3D densely connected convolution neural network with connection-wise attention mechanism for Alzheimer's disease classification

PurposeAlzheimer's disease (AD) is a progressive and irreversible neurodegenerative disease. In recent years, machine learning methods have been widely used on analysis of neuroimage for quantitative evaluation and computer-aided diagnosis of AD or prediction on the conversion from mild cognitive impairment (MCI) to AD. In this study, we aimed to develop a new deep learning method to detect or predict AD in an efficient way.Materials and methodsWe proposed a densely connected convolution neural network with connection-wise attention mechanism to learn the multi-level features of brain MR images for AD classification. We used the densely connected neural network to extract multi-scale features from pre-processed images, and connection-wise attention mechanism was applied to combine connections among features from different layers to hierarchically transform the MR images into more compact high-level features. Furthermore, we extended the convolution operation to 3D to capture the spatial information of MRI. The features extracted from each 3D convolution layer were integrated with features from all preceding layers with different attention, and were finally used for classification. Our method was evaluated on the baseline MRI of 968 subjects from ADNI database to discriminate (1) AD versus healthy subjects, (2) MCI converters versus healthy subjects, and (3) MCI converters versus non-converters.ResultsThe proposed method achieved 97.35% accuracy for distinguishing AD patients from healthy control, 87.82% for MCI converters against healthy control, and 78.79% for MCI converters against non-converters. Compared with some neural networks and methods reported in recent studies, the classification performance of our proposed algorithm was among the top ranks and improved in discriminating MCI subjects who were in high risks of conversion to AD.ConclusionsDeep learning techniques provide a powerful tool to explore minute but intricate characteristics in MR images which may facilitate early diagnosis and prediction of AD.     

基于注意力机制与图卷积神经网络的单目红外图像深度估计

对场景中的物体进行深度估计是无人驾驶领域中的关键问题,红外图像有利于在光线不佳的情况下解决深度估计问题.针对红外图像纹理不清晰与边缘信息不丰富的特点,提出了将注意力机制与图卷积神经网络相结合来解决单目红外图像深度估计问题.首先,在深度估计问题中,图像中每个像素点的深度信息不仅与其周围像素点的深度信息相关,还需考虑更大范...     

多局部残差连接注意网络的图像去模糊

针对现有的基于卷积神经网络的图像去模糊算法存在图像纹理细节恢复不清晰的问题,提出了一种基于多局部残差连接注意网络的图像去模糊算法。首先,采用一个卷积层进行浅层特征提取;其次,设计了一种新的基于残差连接和并行注意机制的多局部残差连接注意模块,用于消除图像模糊并提取上下文信息;再次,采用一个基于扩张卷积的成对连接模块进行细节恢复;最后,利用一个卷积层重建清晰图像。实验结果表明：在GoPro数据集上的PSNR (peak signal to noise ratio)和SSIM (structure similarity)分别为31.83 dB、0.927 5,在定性和定量两方面都表明所提方法能够有效地恢复模糊图像的纹理细节,网络性能优于对比方法。     

多尺度条件卷积的OCT视网膜图像降噪研究

散斑噪声存在于光学相干层析成像(OCT)中,影响OCT图像质量.在使用OCT设备诊断各种常见眼科疾病时,高质量的OCT图像是极为重要的.利用深度神经网络对OCT图像进行降噪处理,使图像在保留空间结构细节的基础上能展示更多的信息.提出了一种基于残差学习网络的新型OCT图像降噪网络-CMCNN,其具有多尺度、多权重和多层次...     

基于级联卷积神经网络的前列腺磁共振图像分类

针对深度学习训练成本高,以及基于磁共振图像的前列腺癌临床诊断需要大量医学常识且极为耗时的问题,本文提出了一种基于级联卷积神经网络（Convolutional Neural Network,CNN）和磁共振图像的前列腺癌（Prostate Cancer,PCa）自动分类诊断方法,该网络以Faster-RCNN作为前网络,对前列腺区域进行提取分割,用于排除前列腺附近组织器官的干扰;以基于ResNet改进的网络结构CNN40_bottleneck作为后网络,用于对前列腺区域病变进行分类.后网络由瓶颈结构串联组成,其中使用批量标准化（Batch Normalization,BN）、全局平均池化（Global Average Pooling,GAP）进行优化.实验结果证明,本文方法对前列腺癌诊断结果较好,而且缩减了训练时间和参数量,有效降低了训练成本.     

Transfer learning in deep neural network based under-sampled MR image reconstruction

In Magnetic Resonance Imaging (MRI), the success of deep learning-based under-sampled MR image reconstruction depends on: (i) size of the training dataset, (ii) generalization capabilities of the trained neural network. Whenever there is a mismatch between the training and testing data, there is a need to retrain the neural network from scratch with thousands of MR images obtained using the same protocol. This may not be possible in MRI as it is costly and time consuming to acquire data. In this research, a transfer learning approach i.e. end-to-end fine tuning is proposed for U-Net to address the data scarcity and generalization problems of deep learning-based MR image reconstruction. First the generalization capabilities of a pre-trained U-Net (initially trained on the human brain images of 1.5 T scanner) are assessed for: (a) MR images acquired from MRI scanners of different magnetic field strengths, (b) MR images of different anatomies and (c) MR images under-sampled by different acceleration factors. Later, end-to-end fine tuning of the pre-trained U-Net is proposed for the reconstruction of the above-mentioned MR images (i.e. (a), (b) and (c)). The results show successful reconstructions obtained from the proposed method as reflected by the Structural SIMilarity index, Root Mean Square Error, Peak Signal-to-Noise Ratio and central line profile of the reconstructed images.     

一种空域和频域相结合的运动图像亚像素配准技术

针对视频图像运动检测问题,提出了一种结合空域灰度投影和频域相位相关的亚像素图像配准方法。首先,采用灰度投影算法在空域对运动图像进行粗配准,即在图像行和列方向上计算图像灰度投影特征数据,根据灰度相关函数最小化准则,估计像素级运动量;然后,在经过粗配准的两幅图像中心选取尺寸相同的区域,进行快速傅里叶变换,在频域采用扩展的相位相关算法对图像进行精确配准。该方法利用图像的功率谱信息,减少对图像内容的依赖,运用基于最小二乘的曲面拟合法,实现亚像素图像配准参数估计,具有从粗到精的特点,有效提高了图像检测精度。文中最后对样本图像进行了图像配准对比实验,结果表明,该方法可以检测0.01 pixel的运动量,最大配准误差为0.004 8 pixel。     

Incorporating prior knowledge via volumetric deep residual network to optimize the reconstruction of sparsely sampled MRI

For sparse sampling that accelerates magnetic resonance (MR) image acquisition, non-linear reconstruction algorithms have been developed, which incorporated patient specific a prior information. More generic a prior information could be acquired via deep learning and utilized for image reconstruction. In this study, we developed a volumetric hierarchical deep residual convolutional neural network, referred to as T-Net, to provide a data-driven end-to-end mapping from sparsely sampled MR images to fully sampled MR images, where cartilage MR images were acquired using an Ultra-short TE sequence and retrospectively undersampled using pseudo-random Cartesian and radial acquisition schemes. The network had a hierarchical architecture that promoted the sparsity of feature maps and increased the receptive field, which were valuable for signal synthesis and artifact suppression. Relatively dense local connections and global shortcuts were established to facilitate residual learning and compensate for details lost in hierarchical processing. Additionally, volumetric processing was adopted to fully exploit spatial continuity in three-dimensional space. Data consistency was further enforced. The network was trained with 336 three-dimensional images (each consisting of 32 slices) and tested by 24 images. The incorporation of a priori information acquired via deep learning facilitated high acceleration factors (as high as 8) while maintaining high image fidelity (quantitatively evaluated using the structural similarity index measurement). The proposed T-Net had an improved performance as compared to several state-of-the-art networks.     

Deriving new soft tissue contrasts from conventional MR images using deep learning

Versatile soft tissue contrast in magnetic resonance imaging is a unique advantage of the imaging modality. However, the versatility is not fully exploited. In this study, we propose a deep learning-based strategy to derive more soft tissue contrasts from conventional MR images obtained in standard clinical MRI. Two types of experiments are performed. First, MR images corresponding to different pulse sequences are predicted from one or more images already acquired. As an example, we predict T_1ρ weighted knee image from T₂ weighted image and/or T₁ weighted image. Furthermore, we estimate images corresponding to alternative imaging parameter values. In a representative case, variable flip angle images are predicted from a single T₁ weighted image, whose accuracy is further validated in quantitative T₁ map subsequently derived. To accomplish these tasks, images are retrospectively collected from 56 subjects, and self-attention convolutional neural network models are trained using 1104 knee images from 46 subjects and tested using 240 images from 10 other subjects. High accuracy has been achieved in resultant qualitative images as well as quantitative T₁ maps. The proposed deep learning method can be broadly applied to obtain more versatile soft tissue contrasts without additional scans or used to normalize MR data that were inconsistently acquired for quantitative analysis.     

High-performance rapid MR parameter mapping using model-based deep adversarial learning

PurposeTo develop and evaluate a deep adversarial learning-based image reconstruction approach for rapid and efficient MR parameter mapping.MethodsThe proposed method provides an image reconstruction framework by combining the end-to-end convolutional neural network (CNN) mapping, adversarial learning, and MR physical models. The CNN performs direct image-to-parameter mapping by transforming a series of undersampled images directly into MR parameter maps. Adversarial learning is used to improve image sharpness and enable better texture restoration during the image-to-parameter conversion. An additional pathway concerning the MR signal model is added between the estimated parameter maps and undersampled k-space data to ensure the data consistency during network training. The proposed framework was evaluated on T₂ mapping of the brain and the knee at an acceleration rate R = 8 and was compared with other state-of-the-art reconstruction methods. Global and regional quantitative assessments were performed to demonstrate the reconstruction performance of the proposed method.ResultsThe proposed adversarial learning approach achieved accurate T₂ mapping up to R = 8 in brain and knee joint image datasets. Compared to conventional reconstruction approaches that exploit image sparsity and low-rankness, the proposed method yielded lower errors and higher similarity to the reference and better image sharpness in the T₂ estimation. The quantitative metrics were normalized root mean square error of 3.6% for brain and 7.3% for knee, structural similarity index of 85.1% for brain and 83.2% for knee, and tenengrad measures of 9.2% for brain and 10.1% for the knee. The adversarial approach also achieved better performance for maintaining greater image texture and sharpness in comparison to the CNN approach without adversarial learning.ConclusionThe proposed framework by incorporating the efficient end-to-end CNN mapping, adversarial learning, and physical model enforced data consistency is a promising approach for rapid and efficient reconstruction of quantitative MR parameters.     

Weakly Supervised Building Semantic Segmentation Based on Spot-Seeds and Refinement Process

Automatic building semantic segmentation is the most critical and relevant task in several geospatial applications. Methods based on convolutional neural networks (CNNs) are mainly used in current building segmentation. The requirement of huge pixel-level labels is a significant obstacle to achieve the semantic segmentation of building by CNNs. In this paper, we propose a novel weakly supervised framework for building segmentation, which generates high-quality pixel-level annotations and optimizes the segmentation network. A superpixel segmentation algorithm can predict a boundary map for training images. Then, Superpixels-CRF built on the superpixel regions is guided by spot seeds to propagate information from spot seeds to unlabeled regions, resulting in high-quality pixel-level annotations. Using these high-quality pixel-level annotations, we can train a more robust segmentation network and predict segmentation maps. To iteratively optimize the segmentation network, the predicted segmentation maps are refined, and the segmentation network are retrained. Comparative experiments demonstrate that the proposed segmentation framework achieves a marked improvement in the building’s segmentation quality while reducing human labeling efforts.     

3D卷积自动编码网络的高光谱异常检测

高光谱图像包含丰富的地物光谱信息,在遥感图像领域有着巨大的发展前景.高光谱图像异常检测无需任何先验光谱信息,便可检测出图像中的异常目标.因此,在国防军事和民用领域都有广泛的应用,是现阶段高光谱图像处理领域的研究热点.然而,高光谱图像存在数据复杂、冗余性强、未标记以及样本数量少等特点,这给高光谱图像异常检测带来了很大的挑...     

A robust algorithm for high-resolution dynamic MRI based on the partially separable functions model

A recently developed partially separable functions (PSF) model can be used to generate high-resolution dynamic magnetic resonance imaging (MRI). However, this method could not robustly reconstruct high-quality MR images because the estimation of the PSF parameters is often interfered by the noise of the sampled MR data. To improve the robustness of MRI reconstruction using the PSF model, we proposed a new algorithm to estimate the PSF parameters by jointly using robust principal component analysis and modified truncated singular value decomposition regularization methods, instead of using the least square fitting method in the original PSF model. The experiment results of in vivo cardiac MRI demonstrated that the proposed algorithm can robustly reconstruct dynamic MR images with higher signal-to-noise ratio and clearer anatomical structures in comparison with the previous PSF model.     

基于小波域三状态HMT模型的含噪图像增强

针对含噪图像增强问题,提出一种基于小波域三状态隐马尔可夫树模型的方法,采用三状态的高斯混合模型逼近小波系数的分布,不需要设定精确的阈值,依据期望最大算法训练得到的每个系数所属状态的后验概率,将系数区分为噪声系数、弱边缘系数和强边缘系数,然后通过抑制噪声系数,增强细节特征系数来达到对含噪图像增强的目的,并引入循环平移策略避免人工失真.通过对含噪的标准图像和人脑核磁共振图像进行仿真实验,并与几种经典的图像增强方法作视觉上的对比和定量分析.实验结果表明,本文所提出的方法具有很好的鲁棒性,在突出了图像中更多的细节信息的同时,可以有效抑制噪声.     

一种被动声呐信号检测的低虚警率方法

针对被动声呐多目标信号检测中的噪声背景归一化问题,提出了一种基于数学形态学滤波的噪声背景归一化新方法。该方法利用数学形态学处理中的膨胀和腐蚀算子,以及基于多项式拟合的数据均值估计方法,构造出了一种能够较为准确的估计噪声门限的方法,并以之进行噪声背景归一化,在较好保留信号信息的前提下较大程度的抑制了噪声,有效降低了多目标信号检测的虚警概率。通过计算机仿真对比了该算法与S3PM算法、OTA算法的性能,结果表明该噪声背景归一化算法能够在检测概率损失较小的情况下较大幅度地降低检测的虚警率。实际被动声呐数据处理的对比结果同样验证了该算法的有效性。      

Automatic segmentation for brain MR images via a convex optimized segmentation and bias field correction coupled model

Accurate segmentation of magnetic resonance (MR) images remains challenging mainly due to the intensity inhomogeneity, which is also commonly known as bias field. Recently active contour models with geometric information constraint have been applied, however, most of them deal with the bias field by using a necessary pre-processing step before segmentation of MR data. This paper presents a novel automatic variational method, which can segment brain MR images meanwhile correcting the bias field when segmenting images with high intensity inhomogeneities. We first define a function for clustering the image pixels in a smaller neighborhood. The cluster centers in this objective function have a multiplicative factor that estimates the bias within the neighborhood. In order to reduce the effect of the noise, the local intensity variations are described by the Gaussian distributions with different means and variances. Then, the objective functions are integrated over the entire domain. In order to obtain the global optimal and make the results independent of the initialization of the algorithm, we reconstructed the energy function to be convex and calculated it by using the Split Bregman theory. A salient advantage of our method is that its result is independent of initialization, which allows robust and fully automated application. Our method is able to estimate the bias of quite general profiles, even in 7T MR images. Moreover, our model can also distinguish regions with similar intensity distribution with different variances. The proposed method has been rigorously validated with images acquired on variety of imaging modalities with promising results.