Image quality assessment using the singular value decomposition theorem

In objective image quality metrics, one of the most important factors is the correlation of their results with the perceived quality measurements. In this paper, a new method is presented based on comparing between the structural properties of the two compared images. Based on the mathematical concept of the singular value decomposition (SVD) theorem, each matrix can be factorized to the products of three matrices, one of them related to the luminance value while the two others show the structural content information of the image. A new method to quantify the quality of images is proposed based on the projected coefficients and the left singular vector matrix of the disturbed image based on the right singular vector matrix of the original image. To evaluate this performance, many tests have been done using a widespread subjective study involving 779 images of the Live Image Quality Assessment Database, Release 2005. The objective results show a high rate of correlation with subjective quality measurements.     

基于奇异值分解的运动模糊图像广义逆恢复方法

当线性模型应用于运动模糊模糊图像的恢复时,方程的最小二乘解是恢复图像的最优线性无偏估计。由于图像退化过程的不适定性,当观测值受到噪声干扰时,该解往往会远偏离真值。为了克服这个问题,通过对退化矩阵的奇异值分解,提取其不易受干扰的子空间,用该空间重构的逆矩阵具有良好抑噪能力,使图像在较长的运动模糊尺度内恢复时保持较低的失真。     

奇异值非线性修正的低对比度红外图像实时增强

由于场景中目标与背景的温差相对较小,红外图像会存在对比度低、视觉效果差的问题,针对这一问题,提出一种基于奇异值非线性修正的红外图像对比度实时增强方法。该方法首先对红外图像进行奇异值分解得到其原始奇异值,然后采用一个对数型非线性变换对图像奇异值进行优化,最后根据修正的奇异值重构出对比度增强的红外图像。利用对数型非线性变换修正图像奇异值不仅能够有效拉伸奇异值的动态范围,同时可优化奇异值的变化梯度,使图像的能量信息得到更充分地表达,改善红外图像不良的视觉效果。实验结果表明,该方法较几种对比方法在视觉效果和客观评价方面均具有更优的增强性能;同时体现出良好的实时性,为实现红外图像的实时增强提供了新途径。     

A distortion compensation method for a conical holographic stereogram

An image processing technique to display a distortion-free reconstructed image from a rainbow type conical holographic stereogram is proposed. The method consists of forming a new set of 2D images from the original set by applying an image process. The image processing method is made through ray tracing equations. The results show that the performance of the proposed method is good as it removes distortion of the image.     

局部方差在图像质量评价中的应用

将灰度图像的局部方差分布(QLS)作为表征图像结构信息的一个重要特征,对局部方差分布矩阵进行奇异值分解,计算得到相应的奇异值特征向量;通过计算降质图像与原参考图像局部方差矩阵奇异值特征向量的夹角大小度量两图像的结构相似度,实现了对降质图像的质量评价。实验结果表明:局部方差分布更能突出图像的结构特征,评价结果优于传统的均方误差(MSE)、峰值信噪比(PSNR)、结构相似度(SSIM)以及直接评价图像像素分布的奇异值分析(SVD)等方法,与人眼视觉感知效果的一致性较好。     

Phase image encryption in the fractional Hartley domain using Arnold transform and singular value decomposition

A novel scheme for image encryption of phase images is proposed, using fractional Hartley transform followed by Arnold transform and singular value decomposition in the frequency domain. Since the plaintext is a phase image, the mask used in the spatial domain is a random amplitude mask. The proposed scheme has been validated for grayscale images and is sensitive to the encryption parameters such as the order of the Arnold transform and the fractional orders of the Hartley transform. We have also evaluated the scheme's resistance to the well-known noise and occlusion attacks.     

基于FPGA的图像畸变矫正算法研究

在机器视觉中需要对采集到的带有畸变的图像进行矫正。为了提高对畸变图像校正的实时性，针对使用FPGA实现图像畸变矫正算法时，存在的在线计算逆向映射坐标复杂和片上ROM容量不够问题，压缩了逆向映射表，并在FPGA上利用插值方法在线重建了逆向映射表。通过查找重建的逆向映射表来获取逆向映射坐标，从而降低了FPGA的在线计算量和片上ROM的容量需求。MATLAB仿真结果显示，当压缩参数n分别取4、8、16时都能够对畸变图像进行较好的矫正，并且图像信息不会丢失。仿真结果验证了算法的有效性，该算法可用于基于FPGA的图像畸变矫正。     

Binary encoding method to encrypt Fourier-transformed information of digital images

An encoding method is used to encrypt the Fourier-transformed information of a hidden (covert) digital image in an overt image, while the Fourier-transformed information must be encoded with binary codes. All of the pixels in an overt image are classified into five groups that are called identification, type, tracing, dimension, and information codes. Identification codes are used to judge if the overt image contains codes that belong to the proposed encoding method or not; type codes are used to judge the encoding type; tracing codes are used to judge the encoding trace; dimension codes are used to judge the size of the hidden information; and information codes are used to decode the hidden information. Applying the proposed encoding method is rather easy, and host images corresponding to overt images are not needed for decoding work. The experiment has demonstrated four types of encoding for the proposed encoding method to reconstruct covert images without any distortion or only with a little distortion.     

A robust image watermarking technique using SVD and differential evolution in DCT domain

In this paper we applied differential evolution (DE) algorithm to balance the tradeoff between robustness and imperceptibility by exploring multiple scaling factors in image watermarking. First of all, the original image is partitioned into blocks and the blocks are transformed into Discrete Cosine Transform (DCT) domain. The DC coefficients from each block are collected to construct a low-resolution approximation image and apply Singular Value Decomposition (SVD) on this approximation image. After that watermark is embedded by modifying singular values with the singular values of the watermark. The role of DE algorithm is to identify the best multiple scaling factors for embedding process in order to achieve the best performance in terms of robustness without compromising with the quality of the image. To enhance the security, watermark is scrambled by Arnold transform before embedding. Experimental results show that the proposed scheme maintains a satisfactory image quality and watermark can still be identified from a seriously distorted image.     

基于小波分解和进化策略的图像融合方法

提出了一种基于小波分解和进化策略相结合的多聚焦图像融合方法。该方法首先对不同聚焦点图像进行无下采小波分解，并分别将其各方向、各尺度的高频信息进行叠加，然后依据它们在高频信息叠加层上对应的局部区域能量的差异，采用进化策略进行区域划分，进而实现分区域融合。实验结果表明所提出的方法比小波变换法具有更好的融合效果。