Fully Learnable Model for Task-Driven Image Compressed Sensing
This study primarily investigates image sensing at low sampling rates with convolutional neural networks (CNN) for specific applications. To improve the image acquisition efficiency in energy-limited systems, this study, inspired by compressed sensing, proposes a fully learnable model for task-drive...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-07-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/21/14/4662 |
_version_ | 1797526089455632384 |
---|---|
author | Bowen Zheng Jianping Zhang Guiling Sun Xiangnan Ren |
author_facet | Bowen Zheng Jianping Zhang Guiling Sun Xiangnan Ren |
author_sort | Bowen Zheng |
collection | DOAJ |
description | This study primarily investigates image sensing at low sampling rates with convolutional neural networks (CNN) for specific applications. To improve the image acquisition efficiency in energy-limited systems, this study, inspired by compressed sensing, proposes a fully learnable model for task-driven image-compressed sensing (FLCS). The FLCS, based on Deep Convolution Generative Adversarial Networks (DCGAN) and Variational Auto-encoder (VAE), divides the image-compressed sensing model into three learnable parts, i.e., the Sampler, the Solver and the Rebuilder. To be specific, a measurement matrix suitable for a type of image is obtained by training the Sampler. The Solver calculates the image’s low-dimensional representation with the measurements. The Rebuilder learns a mapping from the low-dimensional latent space to the image space. All the mentioned could be trained jointly or individually for a range of application scenarios. The pre-trained FLCS reconstructs images with few iterations for task-driven compressed sensing. As indicated from the experimental results, compared with existing approaches, the proposed method could significantly improve the reconstructed images’ quality while decreasing the running time. This study is of great significance for the application of image-compressed sensing at low sampling rates. |
first_indexed | 2024-03-10T09:25:12Z |
format | Article |
id | doaj.art-f7906288a132444fb7232a71e0eff00d |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-10T09:25:12Z |
publishDate | 2021-07-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-f7906288a132444fb7232a71e0eff00d2023-11-22T04:54:11ZengMDPI AGSensors1424-82202021-07-012114466210.3390/s21144662Fully Learnable Model for Task-Driven Image Compressed SensingBowen Zheng0Jianping Zhang1Guiling Sun2Xiangnan Ren3College of Electronic Information and Optical Engineering, Nankai University, Tianjin 300350, ChinaElectrical Engineering and Computer Science, Northwestern University, Evanston, IL 60208, USACollege of Electronic Information and Optical Engineering, Nankai University, Tianjin 300350, ChinaCollege of Electronic Information and Optical Engineering, Nankai University, Tianjin 300350, ChinaThis study primarily investigates image sensing at low sampling rates with convolutional neural networks (CNN) for specific applications. To improve the image acquisition efficiency in energy-limited systems, this study, inspired by compressed sensing, proposes a fully learnable model for task-driven image-compressed sensing (FLCS). The FLCS, based on Deep Convolution Generative Adversarial Networks (DCGAN) and Variational Auto-encoder (VAE), divides the image-compressed sensing model into three learnable parts, i.e., the Sampler, the Solver and the Rebuilder. To be specific, a measurement matrix suitable for a type of image is obtained by training the Sampler. The Solver calculates the image’s low-dimensional representation with the measurements. The Rebuilder learns a mapping from the low-dimensional latent space to the image space. All the mentioned could be trained jointly or individually for a range of application scenarios. The pre-trained FLCS reconstructs images with few iterations for task-driven compressed sensing. As indicated from the experimental results, compared with existing approaches, the proposed method could significantly improve the reconstructed images’ quality while decreasing the running time. This study is of great significance for the application of image-compressed sensing at low sampling rates.https://www.mdpi.com/1424-8220/21/14/4662convolutional neural networkscompressed sensingdeep learningimage reconstruction |
spellingShingle | Bowen Zheng Jianping Zhang Guiling Sun Xiangnan Ren Fully Learnable Model for Task-Driven Image Compressed Sensing Sensors convolutional neural networks compressed sensing deep learning image reconstruction |
title | Fully Learnable Model for Task-Driven Image Compressed Sensing |
title_full | Fully Learnable Model for Task-Driven Image Compressed Sensing |
title_fullStr | Fully Learnable Model for Task-Driven Image Compressed Sensing |
title_full_unstemmed | Fully Learnable Model for Task-Driven Image Compressed Sensing |
title_short | Fully Learnable Model for Task-Driven Image Compressed Sensing |
title_sort | fully learnable model for task driven image compressed sensing |
topic | convolutional neural networks compressed sensing deep learning image reconstruction |
url | https://www.mdpi.com/1424-8220/21/14/4662 |
work_keys_str_mv | AT bowenzheng fullylearnablemodelfortaskdrivenimagecompressedsensing AT jianpingzhang fullylearnablemodelfortaskdrivenimagecompressedsensing AT guilingsun fullylearnablemodelfortaskdrivenimagecompressedsensing AT xiangnanren fullylearnablemodelfortaskdrivenimagecompressedsensing |