一种改进的基于小波方向对比度的无缝拼接算法

为了提高图像无缝拼接效果,对现有平滑算法进行分析,在此基础上提出一种基于小波方向对比度的中缝平滑算法。该算法首先对源图像进行基于统计参数的预处理,使得图像整体亮度趋于一致,然后利用小波变换对图像进行多分辨率分解,对于分解后的图像提出采用一种基于改进的小波方向对比度的规则进行平滑处理。文中还通过几种图像客观评价标准对仿真结果进行了评价,结果表明：该方法在两图像灰度相差较大时也能够取得较好的平滑效果,对光照具有鲁棒性,并且平滑后的图像可以保留原始图像的细节信息,使得图像在保证清晰度的前提下达到无缝拼接。     

A Generative Adversarial Network for Infrared and Visible Image Fusion Based on Semantic Segmentation

This paper proposes a new generative adversarial network for infrared and visible image fusion based on semantic segmentation (SSGAN), which can consider not only the low-level features of infrared and visible images, but also the high-level semantic information. Source images can be divided into foregrounds and backgrounds by semantic masks. The generator with a dual-encoder-single-decoder framework is used to extract the feature of foregrounds and backgrounds by different encoder paths. Moreover, the discriminator’s input image is designed based on semantic segmentation, which is obtained by combining the foregrounds of the infrared images with the backgrounds of the visible images. Consequently, the prominence of thermal targets in the infrared images and texture details in the visible images can be preserved in the fused images simultaneously. Qualitative and quantitative experiments on publicly available datasets demonstrate that the proposed approach can significantly outperform the state-of-the-art methods.     

Approach to Generate Radiometric Images

A new technique to generate radiometric images from the optical images is introduced in this paper. This technique can be applied in generating the source images for purpose of image-matching navigation. For comparison, two radiometric images are presents, one is imaged by a 8mm ground-based radiometer and the other is generated by this technique.     

基于ICT切片图像相邻层信息的缺陷自动识别方法

 现有的工业计算机断层成像（ICT）图像缺陷识方法中,多采用对单张图像进行孤立评判方法,此类方法未能考虑到单张图像在相邻层图像信息关联性,因而易将孤立的噪音视为缺陷,造成误判。为解决这一问题,提出一种基于序列ICT切片图像自动识别方法,该方法将识别过程分为两步：单张图像的潜在缺陷提取和相邻层图像缺陷的匹配。第一步运用传统方法识别出每张图像中所有潜在缺陷;第二步根据真缺陷在相邻层具有匹配关系而伪缺陷则相对孤立的特点,通过分步匹配的方法确定每张图像上所有潜在缺陷在相邻层图像上的匹配关系,区分出真伪缺陷。最后通过实例验证表明：利用该方法可以有效得提高真缺陷得识别率,降低误判率。     

Optical noise reduction by reconstructing positive and negative images from Fourier holograms in coaxial holographic storage systems

A technique for reconstructing positive and negative images from an identical intensity-modulated hologram is demonstrated theoretically and experimentally by use of a coaxial holographic storage system. Negative images are obtained by adding a phase-modulated dc component of signal beam on reading. By comparing positive and negative images, the bit error rate (BER) is improved by two orders of magnitude. This technique can reduce optical noise of reconstructed images to attain low BERs.     

反射式多路投影立体显示系统的研究

设计了一种反射式多路投影立体显示系统,并介绍了该系统的成像原理及特性。系统将12幅视差图同时投影在柱面栅上,观察者仅凭肉眼就可看到具有双眼视差和运动视差的立体图像。系统成像视域较大,可以同时供多人观看。     

Improving the Accuracy of Nearest-Neighbor Classification Using Principled Construction and Stochastic Sampling of Training-Set Centroids

A conceptually simple way to classify images is to directly compare test-set data and training-set data. The accuracy of this approach is limited by the method of comparison used, and by the extent to which the training-set data cover configuration space. Here we show that this coverage can be substantially increased using coarse-graining (replacing groups of images by their centroids) and stochastic sampling (using distinct sets of centroids in combination). We use the MNIST and Fashion-MNIST data sets to show that a principled coarse-graining algorithm can convert training images into fewer image centroids without loss of accuracy of classification of test-set images by nearest-neighbor classification. Distinct batches of centroids can be used in combination as a means of stochastically sampling configuration space, and can classify test-set data more accurately than can the unaltered training set. On the MNIST and Fashion-MNIST data sets this approach converts nearest-neighbor classification from a mid-ranking- to an upper-ranking member of the set of classical machine-learning techniques.     

航天光学遥感器在轨MTF评价的一元方法

通过对航天光学遥感器MTF模型和遥感图片的分析,从图像中提取出与MTF有关的特征信息,采用人工神经网络(ANN)作为工具,将这些特征信息作为ANN的输入向量。在对大量MTF已知的遥感图片进行训练后,ANN可以对未知的遥感图片进行MTF测试。这种方法被称为MTF的一元评价方法,即通过对遥感器传输下来的任意一幅地面景物图像进行MTF的在轨评价,无需在地面铺设特定形状靶标或已知的参考图片。实验结果表明,平均评价误差约为5%,具有很强的抗噪声能力。     

基于鬼像和像差分析的高功率激光装置透镜设计

为了设计能够满足高功率固体激光装置需求的空间滤波器透镜,采用矩阵光学方法,分析光束经过单透镜传输时剩余反射形成各阶鬼像的过程,得到鬼像位置与透镜曲率半径的关系式.该关系式表明,透镜的曲率半径设计可以控制鬼像所在的区域.根据高功率激光装置光路排布的特点简化鬼像分析,利用自行开发的鬼像追迹软件分析了主放大级鬼像分布特点.通过改变透镜的曲率半径和倾斜透镜这两种方式,基于鬼像规避和像差最小化两种原则,最终确定主放大级透镜曲率半径的最优选择为1:3弯月型,凸面朗向光学元件集中的方向.该方法可普遍应用于指导高通量复杂光学系统的透镜设计.     

中药光谱成像图像自适应区域增长分割方法

为了消除背景噪声对药材光谱图像检测结果的干扰,根据中药材光谱图像的特点,设计一种能够自适应对中药材光谱图像进行有效区域(ROI)分割的区域增长算法。该区域增长算法根据药材光谱图像的灰度直方图分布来自动选取种子点和分割阈值,在生长的同时进行连通性分析,生长结束后通过区域填充技术来消除图像中出现的孔洞。实验表明:该方法能够自动、准确地进行ROI分割,分割偏差小于8%,并且能较好地消除噪声的干扰,没有产生无意义的生长区域。     

一种基于工业CT图像的岩心孔隙率计算方法

针对目前基于工业CT图像的岩心孔隙计算中都采用局部自定义区域数据的问题,提出了一种基于岩心图像全部数据的孔隙率计算方法。其中,岩心目标的准确提取是该方法的关键。研究了图像中岩心目标与周边噪音的结构和位置关系,并以此为依据提出了一种工业CT图像的岩心目标提取算法。然后以CT图像中的岩心目标为数据源,通过目标内的孔隙像素与目标像素之比来计算单张图像的岩心孔隙率。而针对岩心的序列CT图像,则通过计算每张图像的孔隙率均值来获得该岩心的孔隙率。最后,以一岩心样本的CT序列图像为例,说明所提方法的准确性和有效性。     

Optical noise reduction for dc-removed coaxial holographic data storage

A method of reconstructing positive and negative images from Fourier holograms recorded without the dc components is demonstrated by use of a coaxial holographic storage system. Reconstructed images are obtained by adding a phase-modulated dc component of the signal beam on reading. Contrast reversal of the reconstructed images can be achieved by reversing the readout reference pattern. This method can realize not only optical noise reduction but also less consumption of the dynamic range of the recording medium, potentially contributing to increasing the number of multiplexed holograms.     

All-optical delay of images using slow light

Two-dimensional images carried by optical pulses (2 ns) are delayed by up to 10 ns in a 10 cm cesium vapor cell. By interfering the delayed images with a local oscillator, the transverse phase and amplitude profiles of the images are shown to be preserved. It is further shown that delayed images can be well preserved even at very low light levels, where each pulse contains on average less than one photon.     

Infrared target tracking in multiple feature pseudo-color image with kernel density estimation

Tracking targets in infrared images is a challenging subject due to the low contrast and severe noise. Kernel density estimation (KDE) with robust performance is one of the well-known tracking algorithms. In essence, tracking targets with KDE algorithm is tracking the statistical features of their pixels by the histograms. The universal KDE which can track any features of targets has not been developed. We propose a strategy which does not need to improve on the KDE algorithm itself, but it can make KDE track other features. We first map the features into the pixel intensity and create the feature images. Then these feature images are used to construct the multiple feature pseudo-color images (MFPCIs). The kernel density estimation algorithm tracks targets in MFPCIs can indirectly track these features. Experiments validate that the performance of tracking targets in MFPCIs outperforms that of tracking them in the original infrared images.     

Multi-scale retinex improvement for nighttime image enhancement

We propose a retinex improvement for nighttime image enhancement. Retinex is often used on images under non-uniform illumination in terms of either color or lightness and has satisfactory results to achieve color constancy and dynamic range compression. Few studies focus retinex on nighttime images, especially those under extreme conditions (i.e., images with over-lighted or extremely under-lighted areas or with noise speckles), on which retinex operation can perform badly. Original multi-scale retinex (MSR) is extremely sensitive to noise speckles that cameras produce in low light areas, and it has unsatisfactory effect on areas with normal or intensive illumination. Moreover, original MSR uses a gain-offset method for prior-to-display treatment and can lead to apparent data loss on nighttime images. This paper replaces the logarithm function in MSR with a customized sigmoid function to minimize data loss, and adapts MSR to nighttime images by merging results from sigmoid-MSR with original images. Experiments show our framework, when applied to nighttime images, can preserve areas with normal or intensive lighting and suppress noise speckles in extreme low light areas.     

一种基于特征点跟踪的电子稳像算法

提出一种基于特征点跟踪的稳像算法。该算法从参考帧图像中提取出一组特征点,然后在后续帧中进行基于Kalman滤波的特征点的跟踪,匹配参考帧图像中特征点的坐标和当前图像中基于Kalman滤波修正后得到的特征点的坐标,并通过仿射模型求出位移量及旋转参数,最后进行反向变换,从而得到稳定的视频图像。实验结果表明:该算法稳像效果好,运算复杂度低,且具有较强的鲁棒性。     

Integral color rainbow hologram based on computer image processing

A new method for obtaining color rainbow holograms of 3D scenes using single wavelength laser is proposed and implemented by combining integral photography, computer graphics and rainbow holography. In the method, three integral photography (IP) images whose colors correspond to three primary colors are calculated by simulating integral photography, and exported on the film. By illuminating the IP images with diffused coherent beam from the opposite side, three images of the 3D scene are reproduced and overlapped on a particular image plane by the three linear micro-lens arrays. The linear micro-lens arrays can be regarded as the slits. A color rainbow hologram can be created after a single exposure by interfering the reproduced 3D images and the reference beam. Furthermore, we demonstrate experimentally the new method for a simple case and the results can be generalized in integral photography.     

Joint reconstruction of multiecho MR images using correlated sparsity

This works addresses the problem of reconstructing multiple T1- or T2-weighted images of the same anatomical cross section from partially sampled K-space data. Previous studies in reconstructing magnetic resonance (MR) images from partial samples of the K-space used compressed sensing (CS) techniques to exploit the spatial correlation of the images (leading to sparsity in wavelet domain). Such techniques can be employed to reconstruct the individual T1- or T2-weighted images. However, in the current context, the different images are not really independent; they are images of the same cross section and, hence, are highly correlated. We exploit the correlation between the images, along with the spatial correlation within the images to achieve better reconstruction results than exploiting spatial correlation only.For individual MR images, CS-based techniques lead to a sparsity-promoting optimization problem in the wavelet domain. In this article, we show that the same framework can be extended to incorporate correlation between images leading to group/row sparsity-promoting optimization. Algorithms for solving such optimization problems have already been developed in the CS literature. We show that significant improvement in reconstruction accuracy can be achieved by considering the correlation between different T1- and T2-weighted images. If the reconstruction accuracy is considered to be constant, our proposed group sparse formulation can yield the same result with 33% less K-space samples compared with simple sparsity-promoting reconstruction. Moreover, the reconstruction time by our proposed method is about two to four times less than the previous method.     

Novel quantum secret image-sharing scheme

In this paper, we propose a novel quantum secret image-sharing scheme which constructs m quantum secret images into m+1 quantum share images. A chaotic image generated by the logistic map is utilized to assist in the construction of quantum share images first. The chaotic image and secret images are expressed as quantum image representation by using the novel enhanced quantum representation. To enhance the confidentiality, quantum secret images are scrambled into disordered images through the Arnold transform. Then the quantum share images are constructed by performing a series of quantum swap operations and quantum controlled-NOT operations. Because all quantum operations are invertible, the original quantum secret images can be reconstructed by performing a series of inverse operations. Theoretical analysis and numerical simulation proved both the security and low computational complexity of the scheme, which has outperformed its classical counterparts. It also provides quantum circuits for sharing and recovery processes.     

A post-processing technique for displaying vessels from routine fast-spin-echo images: MRI-derived angiography.

Fast-spin-echo magnetic resonance (MR) images are routine components of a standard MR brain examination. On these images, blood vessels are visible as black flow void. We report that by applying an enhancement filter to a stack of routine fast-spin-echo MR images, projected angiographic images can be generated. The vascular detail in the projected image is similar to that observed in a phase-contrast image. In addition to its advantage in obtaining vessel information from routine images, the proposed post-processing technique is fast, easy to implement and completely automatic. These images provide additional vessel information that is useful when MR angiography is unavailable or as an aid in planning dedicated MR angiographic studies.