Smoothed Projected Landweber Research Articles

Block compressed sensing image reconstruction via deep learning with smoothed projected Landweber

Compressed sensing (CS) is a signal processing framework for effectively reconstructing signal from a small number of measurements obtained by linear projections of the signal. It is an ongoing challenge for the real-time image reconstruction of the computational imaging, including single pixel imaging based on CS. We built a block-based CS (BCS) image reconstruction framework via a deep learning network with smoothed projected Landweber (SPL). A fully connected network performs both BCS linear sensing and non-linear reconstruction stages, and SPL removes the blocking artifacts due to incorporate Wiener filtering into projected Landweber (PL) method at each iteration. The sensing matrix and nonlinear prediction operator are jointly optimized, and the smoothing filtering is coalesced into the PL framework for eliminating high-frequency oscillatory blocking artifact. Experimental results reveal that the optimized scheme outperforms the approach only based on deep neural network. The reconstruction quality can be improved while being only slightly slower, especially the gain of structural similarity is significantly better than peak signal-to-noise ratio, and the reconstruction image texture details are vivid and natural. At 10% sensing rate, the structural similarity maximum (minimum) gain reaches 0.098 (0.021). The proposed approach is not only far superior to other state-of-the-art CS algorithms in terms of reconstruction time and quality but also comparable with up-to-date deep learning methods.

Fuzzy Adaptive-Sampling Block Compressed Sensing for Wireless Multimedia Sensor Networks.

The transmission of high-volume multimedia content (e.g., images) is challenging for a resource-constrained wireless multimedia sensor network (WMSN) due to energy consumption requirements. Redundant image information can be compressed using traditional compression techniques at the cost of considerable energy consumption. Fortunately, compressed sensing (CS) has been introduced as a low-complexity coding scheme for WMSNs. However, the storage and processing of CS-generated images and measurement matrices require substantial memory. Block compressed sensing (BCS) can mitigate this problem. Nevertheless, allocating a fixed sampling to all blocks is impractical since each block holds different information. Although solutions such as adaptive block compressed sensing (ABCS) exist, they lack robustness across various types of images. As a solution, we propose a holistic WMSN architecture for image transmission that performs well on diverse images by leveraging saliency and standard deviation features. A fuzzy logic system (FLS) is then used to determine the appropriate features when allocating the sampling, and each corresponding block is resized using CS. The combined FLS and BCS algorithms are implemented with smoothed projected Landweber (SPL) reconstruction to determine the convergence speed. The experiments confirm the promising performance of the proposed algorithm compared with that of conventional and state-of-the-art algorithms.

Adaptive Multiscale Block Compressed Sensing of Images based on Gray Level Co- Occurrence Matrix

In conventional block compressed sensing (BCS), the images are divided into small fixed-size blocks sampled at the same sub-rate. The sparsities and high-frequency components of the images are ignored, and the reconstruction qualities of the complex texture images are poor. An adaptive multiscale variant of the block compressed sensing was proposed to reconstruct the texture details of the images. The texture features of the images were obtained from the high-frequency components by the three-level wavelet transform and analyzed on the basis of the gray level co-occurrence matrix. A mathematical model was established to adjust the block sizes of the images automatically and allocate the limited sampling resource adaptively. The smoothed projected Landweber (SPL) was utilized to reconstruct the images. The accuracy of the proposed algorithm was verified by the simulation experiments. Results demonstrate that the texture details of the reconstructed images are abundant. The image edges are also clear, and the blocking artifacts are effectively eliminated. The reconstruction qualities of images, especially the partial images, are considerably improved at different sub-sampling rates. The proposed algorithm achieves a 2.42–3.3 dB gain in reconstruction PSNR for the Barbara image over the original BCS-SPL at a sub-sampling rate of 0.3. No remarkable differences are noted between the reconstructed and original texture blocks in visual sensation. The proposed algorithm provides evidence for the compression and reconstruction of the images with complex texture details.

시간 상관관계를 이용한 분산 압축 비디오 센싱 기법의 복원 화질 개선

가장 간단한 샘플링을 위한 목적으로 SPL (Smoothed Projected Landweber)기법 기반의 움직임 보상 블록 압축센싱 기법이 모든 센싱 프레임들에 대해 분산 압축 비디오 센싱 기술이 적용되는 효과적인 방안으로 연구되어 오고 있다. 그러나 기존의 움직임 보상 블록기반의 압축센싱 기법은 매우 간단하여 복원된 위너-지브 프레임에서 우수한 화질을 제공하지 못하는 한계점이 있다. 본 논문에서는 기존의 움직임 보상 블록기반의 압축센싱 기법을 이용한 위너-지브 프레임에서 우수한 화질을 제공될 수 있도록 알고리즘을 변형한다. 즉, 제안된 알고리즘은 참조 프레임이 연속적인 프레임들에 있어 시간적 상관관계에 기초해서 적응적으로 선택되도록 하는 방법으로 설계된다. 다양한 실험 결과를 통하여 제안한 알고리즘은 기존의 알고리즘에 비해 우수한 화질을 제공할 수 있음을 확인한다. For The Purpose of Pursuing the Simplest Sampling, a Motion Compensated Block Compressed Sensing with Smoothed Projected Landweber (MC-BCS-SPL) has been Studied for an Effective Scheme of Distributed Compressive Video Sensing with all Compressed Sensing (CS) Frames. However, Conventional MC-BCS-SPL Scheme is Very Simple and so it Does not Provide Good Visual Qualities in Reconstructed Wyner-Ziv (WZ) Frames. In this Paper, the Conventional Scheme of MC-BCS-SPL is Modified to Provide Better Visual Qualities in WZ Frames. That is, the Proposed Agorithm is Designed in such a way that the Reference Frame may be Adaptively Selected Based on the Temporal Correlation Between Successive Frames. Several Experimental Results show that the Proposed Algorithm Provides Better Visual Qualities than Conventional Algorithm.

Block Compressed Sensing of Images Using Adaptive Granular Reconstruction

In the framework of block Compressed Sensing (CS), the reconstruction algorithm based on the Smoothed Projected Landweber (SPL) iteration can achieve the better rate-distortion performance with a low computational complexity, especially for using the Principle Components Analysis (PCA) to perform the adaptive hard-thresholding shrinkage. However, during learning the PCA matrix, it affects the reconstruction performance of Landweber iteration to neglect the stationary local structural characteristic of image. To solve the above problem, this paper firstly uses the Granular Computing (GrC) to decompose an image into several granules depending on the structural features of patches. Then, we perform the PCA to learn the sparse representation basis corresponding to each granule. Finally, the hard-thresholding shrinkage is employed to remove the noises in patches. The patches in granule have the stationary local structural characteristic, so that our method can effectively improve the performance of hard-thresholding shrinkage. Experimental results indicate that the reconstructed image by the proposed algorithm has better objective quality when compared with several traditional ones. The edge and texture details in the reconstructed image are better preserved, which guarantees the better visual quality. Besides, our method has still a low computational complexity of reconstruction.

Distributed adaptive compressed video sensing using smoothed projected landweber reconstruction

A novel Compressed-Sensing-based (CS-based) Distributed Video Coding (DVC) system, called Distributed Adaptive Compressed Video Sensing (DISACOS), is proposed in this paper. In this system, the input frames are divided into key frames and non-key frames, which are encoded by block CS sampling. The key frames are encoded as CS measurements at substantially higher rates than the non-key frames and decoded by the Smoothed Projected Landweber (SPL) algorithm using multi-hypothesis predictions. For the non-key frames, a small number of CS measurements are first transmitted to detect blocks having low-quality Side Information (SI) generated by the conventional interpolation or extrapolation at the decoder; then, another group of CS measurements are sampled again upon the decoder's request. To fully utilise the CS measurements, we adaptively allocate these measurements to each block in terms of different edge features. Finally, the residual frame is reconstructed using the SPL algorithm and the decoded non-key frame is simply determined as the sum of the residual frame and the SI. Experimental results have revealed that our CS-based DVC system yields better rate-distortion performance when compared with other schemes.