Entropy-assisted adaptive compressive sensing for energy-efficient visual sensors

Compressive Imaging (CI) is a potential sensing technology for energy-efficient visual sensors, and its rate-distortion performance can be improved by adaptive Compressive Sensing (CS) of image. However, due to the unavailability of original image, it is a challenge for CI-based adaptive CS to extract an effective feature from measurements to evaluate the sparsity of image. In view of that, this paper presents an entropy-assisted adaptive CS system, whose merit is its definition of the sensed entropy without the original image. Based on sensed entropy, each image block is allocated sufficient measuring resources, guaranteeing a cost-effective reconstruction of image. Experimental results show that the proposed entropy-assisted adaptive CS system provides better objective and subjective recovery qualities with a low measuring and recovering complexity.

     

结合图像信号显著性的自适应分块压缩采样

均匀分块压缩感知对图像信号进行压缩采样, 无法有效地分离出重要区域和背景区域。为此,本文提出了一种基于显著性的自适应分块压缩采样方法。根据图像信号的显著性,该方法利用四叉树算法进行自适应图像分块,有效分离出重要区域和背景区域。根据区域块的显著度动态设置观测值数量,重要度区域设置高采样率,背景区域设置低采样率,从而提高重要区域的图像重建质量。实验分析表明,在得到更好的视觉效果的同时,本文算法观测值数量较少,且重构图像的PSNR(峰值信噪比)、MSSIM（平均结构相似性）指标,以及运行时间均优于均匀分块压缩采样算法。     

图像自适应分块的压缩感知采样算法

目的 在图像压缩感知过程中,不管是整体采样还是固定分块采样,都不能充分利用图像的稀疏性,存在采样率与图像重构质量的矛盾。提出了一种基于图像纹理变化的自适应分块感知采样算法ABCS(adaptive block compressed sensing),再结合JPEG量化思想,在不降低图像重构质量的前提下降低采样率,更大地提高压缩比。方法 首先进行图像预分块,计算分析各块纹理复杂度,当图像块纹理复杂度低于相应阈值,选择最佳采样率对各块观测采样,当图像块纹理复杂度高于相应阈值,需再分块,重复上述步骤,达到最小16×16块时停止分块。当最小块的纹理复杂度高于最大阈值采用JPEG量化编码,其他块选择匹配的采样率,以压缩感知方式压缩。结果 ABCS算法与典型的压缩感知重构算法结合并与其原始算法比较,在相近采样率条件下,图像重构质量提高明显,尤其在低采样率下性能更佳,如20%采样率下重构图像PSNR值达到30 dB左右。结论 提出的自适应的分块采样充分利用图像的稀疏分布,提高压缩感知的效率;高复杂纹理块采用JPEG编码处理,避免了重构质量差的缺点,同时减少了重构时间。     

An AK-BRP dictionary learning algorithm for video frame sparse representation in compressed sensing

Sparsifying transform is an important prerequisite in compressed sensing. And it is practically significant to research the fast and efficient signal sparse representation methods. In this paper, we propose an adaptive K-BRP (AK-BRP) dictionary learning algorithm. The bilateral random projection (BRP), a method of low rank approximation, is used to update the dictionary atoms. Furthermore, in the sparse coding stage, an adaptive sparsity constraint is utilized to obtain sparse representation coefficient and helps to improve the efficiency of the dictionary update stage further. Finally, for video frame sparse representation, our adaptive dictionary learning algorithm achieves better performance than K-SVD dictionary learning algorithm in terms of computation cost. And our method produces smaller reconstruction error as well.

     

Self-adaptive sampling rate assignment and image reconstruction via combination of structured sparsity and non-local total variation priors

Compressive sensing (CS) is an emerging approach for acquisition of sparse or compressible signals. For natural images, block compressive sensing (BCS) has been designed to reduce the size of sensing matrix and the complexity of sampling and reconstruction. On the other hand, image blocks with varying structures are too different to share the same sampling rate and sensing matrix. Motivated by this, a novel framework of adaptive acquisition and reconstruction is proposed to assign sampling rate adaptively. The framework contains three aspects. First, a small part of sampling rate is employed to pre-sense each block and a novel approach is proposed to estimate its compressibility only from pre-sensed measurements. Next, two assignment schemes are proposed to assign the other part of the sampling rate adaptively to each block based on its estimated compressibility. A higher sampling rate is assigned to incompressible blocks but a lower one to compressible ones. The sensing matrix is constructed based on the assigned sampling rates. The pre-sensed measurements and the adaptive ones are concatenated to form the final measurements. Finally, it is proposed that the reconstruction is modeled as a multi-objects optimization problem which involves the structured sparsity and the non-local total variation prior together. It is simplified into a 3-stage alternating optimization problem and is solved by an augmented Lagrangian method. Experiments on four categories of real natural images and medicine images demonstrate that the proposed framework captures local and nonlocal structures and outperforms the state-of-the-art methods.     

基于灰度共生矩阵的图像自适应分块压缩感知方法

分块压缩感知的提出很好地弥补了大尺寸图像占用资源多、重构耗时长等不足,但重构后的图像存在明显的块效应。针对现有图像纹理复杂度分析不够准确,导致自适应采样率分配后块效应降低不理想的问题,提出了一种基于灰度共生矩阵的图像自适应分块压缩感知方法。该方法通过共生矩阵分析图像的纹理特性,自适应分配采样率,在总采样率不变的前提下使纹理复杂度高的子块获得较高的采样率,纹理复杂度低的子块获得较低的采样率,并用SAMP(Sparsity Adaptive Matching Pursuit)算法实现重构。仿真结果显示,所提方法能够有效地解决块效应问题,尤其对于局部图像而言,重构图像的画质得到了明显改善。     

基于训练字典的压缩感知光谱稀疏化方法

压缩感知理论利用目标的稀疏特性,能从极少的测量值中重构出目标图像,已成为突破奈奎斯特采样定理,实现超分辨成像的一个极具潜力的研究方向,其应用于对地观测遥感成像的一个核心问题在于面对复杂的地物场景,如何探求有效的稀疏化表达方法。对于具有超高数据量的高光谱成像而言,充分利用波段间丰富的冗余光谱信息,研究有效的光谱稀疏化表达方法更加具有实用价值。首先介绍了压缩感知光谱成像以及光谱稀疏化表达的基本原理,然后利用来自ASTER光谱库的多种类型地物光谱数据构建了一种基于K-SVD方法的训练字典,将其与DCT基、小波基分别作为稀疏基,对于几种典型地物目标进行仿真重构,结果表明:所构建的稀疏字典在采样数较少的情况下明显优于DCT基和小波基,在20%的低采样率时即可近乎完美地重构光谱曲线。     

基于块稀疏度估计的压缩感知自适应重构算法

块稀疏信号是一类具有特殊结构的稀疏信号。针对块稀疏信号块稀疏度未知的情况,提出了一种基于块稀疏度估计的自适应重构算法并将其应用于压缩感知。算法首先对信号的块稀疏度进行初步估计计算得到一个支撑块索引集合的估计值,利用得到的估计值对残差进行初始化;接着对测量矩阵的子块和当前残差进行相关性匹配操作以选取信号的支撑块集合;然后依据正则化原则再次对由相关性匹配操作得到的信号支撑块集合进行筛选;最后通过迭代过程获得信号最终的支撑块集合。仿真实验结果表明,提出的算法与现有的块稀疏信号自适应重构算法比较,具有较好的重构成功概率,且算法的平均运行时间更短。     

Improving the reconstruction efficiency of sparsity adaptive matching pursuit based on the Wilkinson matrix

Sparsity adaptive matching pursuit (SAMP) is a greedy reconstruction algorithm for compressive sensing signals. SAMP reconstructs signals without prior information of sparsity and presents better reconstruction performance for noisy signals compared to other greedy algorithms. However, SAMP still suffers from relatively poor reconstruction quality especially at high compression ratios. In the proposed research, the Wilkinson matrix is used as a sensing matrix to improve the reconstruction quality and to increase the compression ratio of the SAMP technique. Furthermore, the idea of block compressive sensing (BCS) is combined with the SAMP technique to improve the performance of the SAMP technique. Numerous simulations have been conducted to evaluate the proposed BCS-SAMP technique and to compare its results with those of several compressed sensing techniques. Simulation results show that the proposed BCS-SAMP technique improves the reconstruction quality by up to six decibels (dB) relative to the conventional SAMP technique. In addition, the reconstruction quality of the proposed BCS-SAMP is highly comparable to that of iterative techniques. Moreover, the computation time of the proposed BCS-SAMP is less than that of the iterative techniques, especially at lower measurement fractions.     

基于自适应级联的注意力网络的超分辨率重建

深度卷积神经网络显著提升了单图像超分辨率的性能. 通常, 网络越深, 性能越好. 然而加深网络往往会急剧增加参数量和计算负荷, 限制了在资源受限的移动设备上的应用. 提出一个基于轻量级自适应级联的注意力网络的单图像超分辨率方法. 特别地提出了局部像素级注意力模块, 给输入特征的每一个特征通道上的像素点都赋以不同的权值, 从而为重建高质量图像选取更精确的高频信息. 此外, 设计了自适应的级联残差连接, 可以自适应地结合网络产生的层次特征, 能够更好地进行特征重用. 最后, 为了充分利用网络产生的信息, 提出了多尺度全局自适应重建模块. 多尺度全局自适应重建模块使用不同大小的卷积核处理网络在不同深度处产生的信息, 提高了重建质量. 与当前最好的类似方法相比, 该方法的参数量更小, 客观和主观度量显著更好.     

一种改进的稀疏度自适应匹配追踪算法

在信号稀疏度未知的情况下,稀疏度自适应匹配追踪算法(Sparsity Adaptive Matching Pursuit,SAMP)是一种广泛应用的压缩感知重构算法。为了优化SAMP算法的性能,提出了一种改进的稀疏度自适应匹配追踪(Improved Sparsity Adaptive Matching Pursuit,ISAMP)算法。该算法引入广义Dice系数匹配准则,能更准确地从测量矩阵中挑选与残差信号最匹配的原子,利用阈值方法选取预选集,并在迭代过程中采用指数变步长。实验结果表明,在相同的条件下,改进后的算法提高了重构质量和运算速度。     

Self-adaptive Spare Representation of Clustering Structure for Hyperspectral Remote Sensing Image Inpainting

A new algorithm for Hyper\|Spectral Image(HSI)inpaintingbased on self\|adaptive sparse representation of clustering structure is proposed.The method can adaptively select the block size according to the feature of remote sensing image.After the pixel is clustered,each band image of HSI can been sparsely represent according to the dictionary learning algorithm,and achieve HSI inpainting through the sparse approximation.The experimental results show that the sparse coefficients obtained by self\|adaptive spare representation can better represent the HSI and improve the Peak Signal\|to\|Noise Ratio(PSNR) of the image.The method proposed in this paper has important significance and application prospect in remote sensing image application.     

Sparsity estimation based adaptive matching pursuit algorithm

Compared with convex optimization algorithms and combination algorithms, greedy pursuit algorithms can balance operational efficiency and reconstruction precision, so they are widely used in the signal reconstruction step of compressed sensing. However, most existing greedy pursuit algorithms only work well if the signal sparsity is known, and their reconstruction performance is influenced by signal sparsity. To more accurately match the sparsity and obtain better reconstruction performance, we propose a greedy pursuit algorithm, the sparsity estimation based adaptive matching pursuit algorithm, which achieves image reconstruction using a signal sparsity estimation based on the Restricted Isometry Property (RIP) criterion and a flexible step size. Experimental results demonstrate that this algorithm provides better reconstruction performance and lower computation time, using different measurement matrices, when the sparsity is estimated in advance.     

使用二分图邻接矩阵的压缩传感图像快速重建

针对压缩传感中高维投影计算采用稀疏性较差的普通随机测量矩阵,从而导致计算复杂度高,重构性能不佳这一难题,提出一种基于二分图邻接矩阵的压缩传感图像快速重建算法。该算法在满足测量矩阵的RIP条件下,充分利用二分图邻接矩阵的稀疏性与二值性,将时间复杂度由传统算法的O(N·logN)降低至O(N)。实验结果表明,算法在保证图像重构质量情况下大大提高了运算性能,尤其对于色彩(灰度)变化平缓图像,该算法性能更加优越。     

DLSLA 3-D SAR imaging algorithm for off-grid targets based on pseudo-polar formatting and atomic norm minimization基于伪极坐标变换和原子范数最小化的网格偏离目标DLSLA 3-D SAR成像方法

This paper concerns the imaging problem for downward looking sparse linear array three-dimensional synthetic aperture radar (DLSLA 3-D SAR) under the circumstance of sparse and non-uniform cross-track dimensional virtual phase centers configuration. Since the 3-D imaging scene behaves typical sparsity in a certain domain, sparse recovery approaches hold the potential to achieve a better reconstruction performance. However, most of the existing compressive sensing (CS) algorithms assume the scatterers located on the pre-discretized grids, which is often violated by the off-grid effect. By contrast, atomic norm minimization (ANM) deals with sparse recovery problem directly on continuous space instead of discrete grids. This paper firstly analyzes the off-grid effect in DLSLA 3-D SAR sparse image reconstruction, and then introduces an imaging method applied to off-gird targets reconstruction which combines 3-D pseudo-polar formatting algorithm (pseudo-PFA) with ANM. With the proposed method, wave propagation and along-track image reconstruction are operated with pseudo-PFA, then the cross-track reconstruction is implemented with semidefinite programming (SDP) based on the ANM model. The proposed method holds the advantage of avoiding the off-grid effect and managing to locate the off-grid targets to accurate locations in different imaging scenes. The performance of the proposed method is verified and evaluated by the 3-D image reconstruction of different scenes, i.e., point targets and distributed scene.     

基于物联网技术的小波域分块压缩感知算法的图像重构系统设计

针对低照明度重构图像分辨率不高、重构时间长的问题,提出了基于小波域分块压缩感知算法的图像重构系统。建立低照明度图像采样模型,采用图像的景深自适应调节方法进行小波域分块压缩感知和信息融合处理。利用多尺度的Retinex算法进行小波域分块压缩感知和信息提取,提取图像的信息熵特征量。采取图像自适应增强方法进行低照度图像增强处理,使用物联网技术进行低照明度图像的三维信息重构,结合细节增强方法进行低照度图像增强处理,完成重构系统设计,实现透射率图的轮廓检测和特征重构。仿真结果表明,采用该方法进行低照明度图像重构的分辨率较高,边缘感知能力较好,且重构耗时较短,实际应用效率较高。     

优化加权TV的复合正则化压缩感知图像重建

目的：压缩感知理论突破了传统的Shanon-Nyquist采样定理的限制,能够以较少的采样值来进行原信号的恢复。针对压缩感知图像重建问题,本文提出了一种基于优化加权全变差（Total Variation, TV）的复合正则化压缩感知图像重建模型。方法：提出的重建模型是以TV正则化模型为基础的。首先,为克服传统TV正则化会导致重建图像的边缘和纹理细节部分模糊或丢失的缺点,本文引入图像的梯度信息估计权重,构建加权TV的重建模型。其次,利用全变差去噪（Rudin–Osher–Fatemi,ROF）模型对权重进行优化估计,从而减少计算权重时受噪声的影响。再次,本文将非局部结构相似性先验和局部自回归性先验引入提出的加权TV模型,得到优化加权TV的复合正则化重建模型。最后,结合投影法和算子分裂法对优化模型求解。结果：针对自然图像的不同特性,本文使用复合正则化先验进行建模,实验表明上述重建问题通过我们的方法得到了很好的解决,加权TV正则化先验使得图像的平坦区域和强边重建较好,而非局部结构相似性先验和局部自回归性先验能够保证图像的精细结构部分的重建效果。结论：本文提出了一种新的复合正则化压缩感知重建模型。与其它基于TV正则化的重建模型相比,实验结果表明本文模型的重建性能无论是在视觉效果还是在客观评价指标上都有明显的提高。     

用EMD和小波消噪的加速度信号压缩重构新方法

针对噪声破坏加速度信号稀疏性、降低其压缩感知重构算法性能问题,提出了一种用经验模态分解(EMD)和小波分析联合消噪的加速度信号压缩重构新方法.该方法首先采用EMD和小波阈值联合消噪方法对加速度信号消噪处理,保持加速度信号内在稀疏性;然后基于压缩感知理论和加速度信号块结构信息,采用块稀疏贝叶斯学习算法以高概率重构原始加速度信号.采用USC-HAD人体日常行为数据库中的加速度信号验证本文方法的有效性.实验结果表明,本文所提方法的信噪比和均方根误差明显优于未经消噪处理的压缩感知重构算法,能够有效抑制加速度信号噪声,增大加速度信号稀疏度,提高加速度信号压缩重构算法性能.     

一种自适应正则化子空间追踪算法

针对压缩感知中未知稀疏度信号的重建问题,提出一种新的压缩感知的信号重建算法,即自适应正则化子空间追踪（Adaptive Regularized Subspace Pursuit,ARSP）算法,该算法将自适应思想、正则化思想与子空间追踪（Subspace Pursuit,SP）算法相结合,在未知信号稀疏度的情况下,自适应地选择支撑集原子的个数,利用正则化过程实现支撑集的二次筛选,最终能实现信号的精确重构。仿真结果表明,该算法能够精确重构原始信号,重建效果优于SP算法、正则化正交匹配追踪（ROMP）算法、稀疏度自适应匹配追踪（SAMP）算法、压缩采样匹配追踪（CoSaMP）算法等。