基于自适应字典选择的MCA图像修复方法

形态成分分析是一种基于稀疏模型的图像分析算法,其中心思想是根据信号组成成分的形态差异性选择两个合适的字典分别用来表示纹理部分和边缘卡通部分,具有良好的图像修复特性。传统上字典的选择需要由使用者根据图像内容人为确定。提出一种基于图像内容的自适应字典选择方法,根据最小能量在字典集合中选择最适合当前图像的字典并对图像进行修复。实验证明,该方法具有良好的图像修复性能。     

Robust object detection based on deformable part model and improved scale invariant feature transform

We propose an approach to improve the detection results of a generic offline trained detector on frames from a specific video. For two consecutive frames of a video with the object, deformable part model (DPM) detection is performed to get the original detections. Then the image patches corresponding to the detected root box and part boxes were respectively obtained. Thirdly, improved scale invariant feature transform features (SIFT) from those image patches were extracted and matched with the SIFT features by KD-Tree. K-means clustering the angle and scale of matched keypoints to filter out the uncorrected matches and further remove false matches by RANSAC algorithm. Finally, the SIFT_DPM detection result from the matches between image patches of continuous frames was obtained. We focus on methods with high precision detection results since it is necessitated in real application. Extensive experiments with state-of-the-art detector demonstrate the efficacy of our approach.     

Planar infrared binary phase reflectarray

A reflective, binary phase reflectarray is demonstrated in the infrared, at a wavelength of 10.6 microm. The unique aspect of this work, at this frequency band, is that the specific desired phase shift is achieved using an array of subwavelength metallic patches on top of a ground-plane-backed dielectric stand-off layer. This is an alternative to the usual method of constructing a reflective Fresnel zone plate by means of a given thickness of dielectric. This initial demonstration of the reflectarray approach at infrared is significant in that there is inherent flexibility to create a range of phase shifts by varying the dimensions of the patches. This will allow for a multilevel phase distribution, or even a continuous variation of phase, across an optical surface with only two-dimensional lithography, avoiding the need for dielectric height variations.     

Learning-based compressed sensing for infrared image super resolution

This paper presents an infrared image super-resolution method based on compressed sensing (CS). First, the reconstruction model under the CS framework is established and a Toeplitz matrix is selected as the sensing matrix. Compared with traditional learning-based methods, the proposed method uses a set of sub-dictionaries instead of two coupled dictionaries to recover high resolution (HR) images. And Toeplitz sensing matrix allows the proposed method time-efficient. Second, all training samples are divided into several feature spaces by using the proposed adaptive k-means classification method, which is more accurate than the standard k-means method. On the basis of this approach, a complex nonlinear mapping from the HR space to low resolution (LR) space can be converted into several compact linear mappings. Finally, the relationships between HR and LR image patches can be obtained by multi-sub-dictionaries and HR infrared images are reconstructed by the input LR images and multi-sub-dictionaries. The experimental results show that the proposed method is quantitatively and qualitatively more effective than other state-of-the-art methods.     

Generation of a uniform line source using a binary filter

A uniform or nearly uniform 10 mm long line source has been generated from a gaussian laser beam using a binary filter. This is done by processing the binary filter throughout a well corrected optical system, so that the equivalent of a continuous inverse gaussian filter is obtained at the final image plane. This equivalent continuous filter generates a uniform line source with maximum conversion efficiency of 50%.     

Blind image deconvolution using sparse and redundant representation

This work devotes to the image deconvolution problem that restores clear image from its blurred and noisy measurements with little prior about the blur. A deconvolution method based on sparse and redundant representation theory is developed in this paper. It firstly represents the blur and image over different redundant dictionaries and imposes sparsity constraint to their representation coefficients respectively, then alternately estimates them using an iterative algorithm employing optimization technique. Experimental results on astronomical images show that the proposed method can achieve as good performance as the method requiring a known blur, which demonstrates its effectiveness.     

Bayes-optimal solution to inverse halftoning based on statistical mechanics of the Q-Ising model

On the basis of statistical mechanics of the Q-Ising model, we formulate the Bayesian inference to the problem of inverse halftoning, which is the inverse process of representing gray-scales in images by means of black and white dots. Using Monte Carlo simulations, we investigate statistical properties of the inverse process, especially, we reveal the condition of the Bayes-optimal solution for which the mean-square error takes its minimum. The numerical result is qualitatively confirmed by analysis of the infinite-range model. As demonstrations of our approach, we apply the method to retrieve a grayscale image, such as standard image Lena, from the halftoned version. We find that the Bayes-optimal solution gives a fine restored grayscale image which is very close to the original. In addition, based on statistical mechanics of the Q-Ising model, we are sucessful in constructing a practically useful method of inverse halftoning using the Bethe approximation.      

制作连续沟形微光学元件的新方法

提出一种制作连续沟形微光学元件的新方法，用移动物面上二元图案的方法来获得连续灰度的记录，再用卤化银明胶处理技术，把灰度银像转化成明胶硬度潜像，最后采用酶蚀显影工艺把潜像显现为连续沟形的浮雕结构，文中给出了原理分析和实验验证，初步结果令人满意。     

三维光滑接触界面上的分配参数滑移面算法

针对三维滑移界面上的分配参数滑移算法,提出一种三维滑移面光滑目的 ,仅根据单元面顶点的坐标和法向等局部信息,构造出曲面片边界G¹连续(切平面连续),曲面片内C¹连续的三维光滑曲面,此目的 为实际的滑移面提供一个准确的几何表达.在此光滑曲面上构建了具有C¹连续的法向压力场和质量面密度场.然后在光滑滑移界面上通过Newton-Raphson迭代计算接触点,依据分配参数滑移算法施加接触约束和接触计算.数值算例显示光滑的接触滑移面可以减少滑移节点的振动,提高计算的收敛性.     

Saliency detection via two-directional 2DPCA analysis of image patches

Salient object detection is an important and challenging problem in computer vision. In this paper, we present a model of salient region detection based on the fusion of contrast and distribution, computed by two-directional 2DPCA analysis of image patches under the combination of RGB space, LAB space and YCbCr space. First, non-overlap patches of three layers from the image are obtained in the three color spaces respectively and stacked for the combination of the three sapces in a single layer. For every layer, two-directional, two-dimensional PCA are utilized to realize automatic selection of effective features, then based on the high contrast and compact character of salient object, contrast values and distribution values of image patches are fused to get the saliency map. Finally, three saliency maps for three layers are combined to detect salient object. The experimental results on a publicly available database show that the proposed algorithm performs well and are in line with the human eye observation results.     

Encrypting image by assembling the fractal-image addition method and the binary encoding method

The fractal-image addition method and the binary encoding method are assembled to form a hybrid method for encrypting a digital covert image. For this hybrid method, a host image is used to create an overt image with the information of the covert image. First, the fractal-image addition method is used to add some fractal images and the covert image to form an image-mixing matrix. Then, all the pixel values of the image-mixing matrix are transferred into binary data. Finally, the binary data are encoded into the host image to create an overt image. The pixels of the overt image contain eight groups of codes used for reconstructing the covert image. The eight groups of codes are identification codes, row amount codes, covert-image dimension codes, fractal-image amount codes, starting-pixel codes, character amount codes, character codes, and information codes. The overt image and the host image look almost the same for eyes. Furthermore, the covert image can be directly reconstructed from the overt image without using the host image. The most important feature is that the reconstructed covert image is identical to the original covert image, i.e. there is no distortion in the decoding work.     

Two-stage non-local means filtering with adaptive smoothing parameter

Non-local means (NLM) filtering is an efficacious algorithm in image denoising which searches the similar neighborhoods and estimates the pixel by averaging these neighborhoods. Some internal parameters such as patch size, search window size and smoothing strength have serious effects on filtering performance. This paper proposes an improved version of NLM by using weak textured patches based single image noise estimation and two-stage NLM with adaptive smoothing parameter. Our proposed method firstly applies weak textured patches based noise estimation to achieve the noise level of input noisy image. Then relying on the estimated noise level, we apply the first stage NLM with adaptive smoothing parameter to attain a basic denoised image. After that, the basic denoised image is refined by the second stage of NLM with smaller smoothing strength. Our experimental results show that the proposed algorithm outperforms the NLM and some NLM recent variants both in visual quality and numerical measures. Additionally, the potential halo effect is almost eliminated in the result images produced by our proposed method.     

提高二元光学器件衍射效率的准连续编码策略

基于标量衍射理论，二元光学器件的衍射效率，微结构误差对衍射效率的影响，提出了一种准连续编码方法，并给出了加工非２ｎ台阶所需掩模板的具体形式。在一定的加工条件和工艺水平下，采用这种编码方法，可以不增加加工成本和工序，仅通过新颖的掩模板设计方法充分利用现有加工条件和设备，使二元光学器件的衍射效率得到很大程度的提高。从而不仅能使一般混合系统的效率得到提高，更能充分发挥混合光学系统的成像能力。此编码方法的采用有利于推进二元光学器件的实用化。     

Gray-Level Image Encryption Scheme Using Full Phase Encryption and Phase-Encoded Exclusive-OR Operations

A new gray-level image encryption scheme based on phase-encoded exclusive-OR (XOR) operations and a full phaseencoding method is presented. The gray-level image can be sliced into binary images which have the same pixel level, and each of them is encrypted by phase-encoded XOR operations with each phase-encoded binary random image. We combine XORed images and phase-encoded binary random images into an encrypted image and a key image, respectively. Gray-level encrypted data and key data are obtained by the phase-encoding process of the encrypted image and the key image in a space domain. The decryption process is simply implemented by using a phasevisualization system and the performance of the proposed scheme is evaluated by computer simulation.     

基于混沌与位平面的图像置乱方法

根据混沌序列良好的随机性及灰度图像可分成若干位平面二值图的特点,提出了将混沌二值序列与位平面图像异或运算进行置乱,将置乱后的位平面图重新组合成灰度图,以此达到灰度图像置乱的效果。仿真结果表明,基于混沌与位平面的图像置乱方法可有效地使图像灰度值均匀化。     

一种基于游程码的二值图像孔洞标记和表达方法

在考察多种二值图像连通区域标记方法的基础上,提出了一种新的基于图像填充和游程码的二值图像孔洞连通域的标记和表达方法。对二值图像进行一次扫描,得到所有可能构成孔洞的游程码,并将游程对应的区域基于特定标记值进行填充。通过判断每个游程与图像上一行像素点标记值的对应关系完成新连通域的建立、游程数据存储和连通区域的合并,判断每个游程与图像下一行像素点标记值的对应关系完成孔洞的筛选。由于仅需扫描图像一次,且对每个游程只须进行单独处理,标记算法的效率较高、复杂度较低。结果表明,可以快速有效地完成二值图像孔洞连通域的标记。同时,它还给出孔洞之间的包含关系,能够有效地实现二值图像的拓扑结构表达。     

Generation of hollow beams by using a binary spatial light modulator

We demonstrate the generation of hollow laser beams by using a binary spatial light modulator and compare the results with those for a continuous modulator. The binary phase modulator produces beams that have continuous, azimuthally varying phase profiles and can be dynamically changed with kilohertz refresh rates. The intensity and phase profiles are recorded through the focus of an imaging lens and are compared with scalar diffraction theory. We highlight properties of the beams relevant to optical dipole traps.     

一种基于保守混沌的密钥分发协议及图像加密算法

耗散混沌系统可以通过时滞嵌入法重构混沌吸引子,因而耗散混沌在基于混沌的信息加密技术中存在一定隐患。针对这一问题提出一种基于保守混沌的密钥分发协议及图像加密算法,该算法将图像数据通过Hash算法转换为保守混沌系统的初始值,形成一次一密的加密结构。然后利用保守混沌信号结合密钥分发协议生成二进制密钥流,该过程由发送方和接受方双方共同完成。使用得到的二进制密钥流对图像数据置乱和扩散,得到加密图像数据。仿真结果表明：该算法安全可靠,能够有效地隐藏原图像信息,并能抵御一些常见的攻击。此外,该加密算法所使用的保守混沌不存在吸引子难以被破解,算法的安全性进一步加强。     

Distortion-Invariant Binary Image Recognition Based on Central Projection Correlation

A method of central projection correlation which is invariant to distortion of shift, scale and rotation of the binary target image is proposed. A 2-D binary image is transformed into an 1-D central projection referring to the centroid of the binary image. The distortion-invariant central projection correlation is successfully performed by computer simulations and its optical implementation is presented.     

Distortion Invariant Binary Image Recognition Based on Central Projection Correlation

1　Ｉｎｔｒｏｄｕｃｔｉｏｎ　　ＴｈｅｍａｔｃｈｅｄｓｐａｔｉａｌｆｉｌｔｅｒｉｎｇｐｒｏｐｏｓｅｄｂｙＶａｎｄｅｒｌｕｇｔ[1] ｈａｓｒｅｃｅｉｖｅｄｃｏｎｓｉｄｅｒａｂｌｅａｔｔｅｎｔｉｏｎｂｅｃａｕｓｅｉｔｈａｓｉｎｈｅｒｅｎｔａｄｖａｎｔａｇｅｏｆｓｈｉｆｔｉｎｖａｒｉａｎｃｅ ,ｂｕｔｔｈｉｓｔｅｃｈｎｉｑｕｅｓｕｆｆｅｒｓｆｒｏｍｓｅｎｓｉｔｉｖｉｔｙｔｏｒｏｔａｔｉｏｎａｎｄｓｃａｌｅｄｅｆｏｒｍａｔｉｏｎ .Ｔｈｅｉｎｖａｒｉａｎｃｅｉｓｔｈｅｋｅｙｔｏｐａｔｔｅｒｎｒｅｃｏｇｎｉｔｉ…