Sequential Convolution and Runge-Kutta Residual Architecture for Image Compressed Sensing

Zheng, Runkai; Zhang, Yinqi; Huang, Daolang; Chen, Qingliang

doi:10.1007/978-3-030-58545-7_14

Cited by 5 publications

(3 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because the fixed measuring matrix used in the CS methods needs a lot of storage space and must satisfy the RIP principle, we used sequential three convolutional layers to replace the fixed measuring matrix in the traditional CS compression process to compress the original signals. The key point is that the convolution can be represented by a multiplication between matrix and matrix [ 29 ]. The process can be formulated as:

where

and

( i = 1, 2, 3) are the weights and bias, respectively, in i -th convolutional layer in the compression module, and y is the output (i.e., measurement).…”

Section: Methodsmentioning

confidence: 99%

Deep Compressive Sensing on ECG Signals with Modified Inception Block and LSTM

Hua

Rao

Peng

et al. 2022

Entropy

View full text Add to dashboard Cite

In practical electrocardiogram (ECG) monitoring, there are some challenges in reducing the data burden and energy costs. Therefore, compressed sensing (CS) which can conduct under-sampling and reconstruction at the same time is adopted in the ECG monitoring application. Recently, deep learning used in CS methods improves the reconstruction performance significantly and can removes of some of the constraints in traditional CS. In this paper, we propose a deep compressive-sensing scheme for ECG signals, based on modified-Inception block and long short-term memory (LSTM). The framework is comprised of four modules: preprocessing; compression; initial; and final reconstruction. We adaptively compressed the normalized ECG signals, sequentially using three convolutional layers, and reconstructed the signals with a modified Inception block and LSTM. We conducted our experiments on the MIT-BIH Arrhythmia Database and Non-Invasive Fetal ECG Arrhythmia Database to validate the robustness of our model, adopting Signal-to-Noise Ratio (SNR) and percentage Root-mean-square Difference (PRD) as the evaluation metrics. The PRD of our scheme was the lowest and the SNR was the highest at all of the sensing rates in our experiments on both of the databases, and when the sensing rate was higher than 0.5, the PRD was lower than 2%, showing significant improvement in reconstruction performance compared to the comparative methods. Our method also showed good recovering quality in the noisy data.

show abstract

where

and

( i = 1, 2, 3) are the weights and bias, respectively, in i -th convolutional layer in the compression module, and y is the output (i.e., measurement).…”

Section: Methodsmentioning

confidence: 99%

Deep Compressive Sensing on ECG Signals with Modified Inception Block and LSTM

Hua

Rao

Peng

et al. 2022

Entropy

View full text Add to dashboard Cite

show abstract

“…We used 4 sequential convolutional layers and 1 up-sampling layer instead of the measurement matrix in the traditional CS method. Because the convolution is representable as a matrix-to-matrix multiplication [ 36 ], a sequential 4-layer convolution formula is (5):

where

and

are the weight and bias in the ith convolutional layer in the CS-Block, respectively. The convolution can be expressed as a linear representation of the original signal

, so convolution can replace the traditional CS sampling matrix in the network model to sample and compress the information.…”

Section: Methodsmentioning

confidence: 99%

Leaf Classification for Crop Pests and Diseases in the Compressed Domain

Hua

Zhu

Liu

2022

Sensors

View full text Add to dashboard Cite

Crop pests and diseases have been the main cause of reduced food production and have seriously affected food security. Therefore, it is very urgent and important to solve the pest problem efficiently and accurately. While traditional neural networks require complete processing of data when processing data, by compressed sensing, only one part of the data needs to be processed, which greatly reduces the amount of data processed by the network. In this paper, a combination of compressed perception and neural networks is used to classify and identify pest images in the compressed domain. A network model for compressed sampling and classification, CSBNet, is proposed to enable compression in neural networks instead of the sensing matrix in conventional compressed sensing (CS). Unlike traditional compressed perception, no reduction is performed to reconstruct the image, but recognition is performed directly in the compressed region, while an attention mechanism is added to enhance feature strength. The experiments in this paper were conducted on different datasets with various sampling rates separately, and our model was substantially less accurate than the other models in terms of trainable parameters, reaching a maximum accuracy of 96.32%, which is higher than the 93.01%, 83.58%, and 87.75% of the other models at a sampling rate of 0.7.

show abstract

“…Recently, deep compressive sensing (DCS) methods (Sun et al 2020;Chen et al 2020;You et al 2021;Song, Chen, and Zhang 2021) have been developed to solve these two issues of CS through an end-to-end learning manner, leveraging the robust learning and representation abilities of neural networks. Zheng et al (2020) introduced RK-CCSNet, a method that employs sequential convolution modules (SCM) to compress image size by means of filter compression, This approach effectively avoids block To address aforementioned challenges, we propose a multi-level cross-sampling and frequency-divided reconstruction network (MCFD-Net) to achieve higher quality image CS, as illustrated in Fig. 2.…”

Section: Introductionmentioning

confidence: 99%

Multi-Cross Sampling and Frequency-Division Reconstruction for Image Compressed Sensing

Song,

Gong,

Meng

et al. 2024

AAAI

View full text Add to dashboard Cite

Deep Compressed Sensing (DCS) has attracted considerable interest due to its superior quality and speed compared to traditional CS algorithms. However, current approaches employ simplistic convolutional downsampling to acquire measurements, making it difficult to retain high-level features of the original signal for better image reconstruction. Furthermore, these approaches often overlook the presence of both high- and low-frequency information within the network, despite their critical role in achieving high-quality reconstruction. To address these challenges, we propose a novel Multi-Cross Sampling and Frequency Division Network (MCFD-Net) for image CS. The Dynamic Multi-Cross Sampling (DMCS) module, a sampling network of MCFD-Net, incorporates pyramid cross convolution and dual-branch sampling with multi-level pooling. Additionally, it introduces an attention mechanism between perception blocks to enhance adaptive learning effects. In the second deep reconstruction stage, we design a Frequency Division Reconstruction Module (FDRM). This module employs a discrete wavelet transform to extract high- and low-frequency information from images. It then applies multi-scale convolution and self-similarity attention compensation separately to both types of information before merging the output reconstruction results. The MCFD-Net integrates the DMCS and FDRM to construct an end-to-end learning network. Extensive CS experiments conducted on multiple benchmark datasets demonstrate that our MCFD-Net outperforms state-of-the-art approaches, while also exhibiting superior noise robustness.

show abstract

Sequential Convolution and Runge-Kutta Residual Architecture for Image Compressed Sensing

Cited by 5 publications

References 27 publications

Deep Compressive Sensing on ECG Signals with Modified Inception Block and LSTM

Deep Compressive Sensing on ECG Signals with Modified Inception Block and LSTM

Leaf Classification for Crop Pests and Diseases in the Compressed Domain

Multi-Cross Sampling and Frequency-Division Reconstruction for Image Compressed Sensing

Contact Info

Product

Resources

About