LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network
The removal of impulse noise is a crucial pre-processing step in image processing systems. In recent years, numerous noise-removal methods have been proposed to improve denoizing performance and reconstruct noise-free images. However, removing high-density impulse noise remains a major challenge. In...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Open Journal of the Computer Society |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9152117/ |
| _version_ | 1818743283200819200 |
|---|---|
| author | Trung-Hieu Le Po-Hsiung Lin Shih-Chia Huang |
| author_facet | Trung-Hieu Le Po-Hsiung Lin Shih-Chia Huang |
| author_sort | Trung-Hieu Le |
| collection | DOAJ |
| description | The removal of impulse noise is a crucial pre-processing step in image processing systems. In recent years, numerous noise-removal methods have been proposed to improve denoizing performance and reconstruct noise-free images. However, removing high-density impulse noise remains a major challenge. In this paper, to address the image denoizing problem associated with high-density noise, we propose a new denoizing model, called LD-Net, which can be trained end-to-end and directly reconstructs noise-free images via a lightweight convolutional neural network. LD-Net is performed in two stages including a feature augmentation stage and a feature refinement stage. During the feature augmentation stage, the spatial size and dimension of the input image are increased by employing the deconvolutional layers for effective feature learning. During the feature refinement stage, the textural details of the image are enhanced for the reconstruction of the noise-free image by the utilization of a proposed sequence of three convolutional layers. Quantitative and qualitative evaluations performed on the SN-LABELME dataset indicate that the proposed LD-Net removes high-density impulse noise more effectively and at higher speed than other state-of-the-art denoizing methods. |
| first_indexed | 2024-12-18T02:25:57Z |
| format | Article |
| id | doaj.art-00e4f210134641bc9874a254022caba7 |
| institution | Directory Open Access Journal |
| issn | 2644-1268 |
| language | English |
| last_indexed | 2024-12-18T02:25:57Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of the Computer Society |
| spelling | doaj.art-00e4f210134641bc9874a254022caba72022-12-21T21:24:02ZengIEEEIEEE Open Journal of the Computer Society2644-12682020-01-01117318110.1109/OJCS.2020.30127579152117LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural NetworkTrung-Hieu Le0https://orcid.org/0000-0001-5766-4199Po-Hsiung Lin1Shih-Chia Huang2https://orcid.org/0000-0002-6896-3415Department of Electronic Engineering, National Taipei University of Technology, Taipei, TaiwanDepartment of Electronic Engineering, National Taipei University of Technology, Taipei, TaiwanDepartment of Electronic Engineering, National Taipei University of Technology, Taipei, TaiwanThe removal of impulse noise is a crucial pre-processing step in image processing systems. In recent years, numerous noise-removal methods have been proposed to improve denoizing performance and reconstruct noise-free images. However, removing high-density impulse noise remains a major challenge. In this paper, to address the image denoizing problem associated with high-density noise, we propose a new denoizing model, called LD-Net, which can be trained end-to-end and directly reconstructs noise-free images via a lightweight convolutional neural network. LD-Net is performed in two stages including a feature augmentation stage and a feature refinement stage. During the feature augmentation stage, the spatial size and dimension of the input image are increased by employing the deconvolutional layers for effective feature learning. During the feature refinement stage, the textural details of the image are enhanced for the reconstruction of the noise-free image by the utilization of a proposed sequence of three convolutional layers. Quantitative and qualitative evaluations performed on the SN-LABELME dataset indicate that the proposed LD-Net removes high-density impulse noise more effectively and at higher speed than other state-of-the-art denoizing methods.https://ieeexplore.ieee.org/document/9152117/LD-Netlightweight denoising modelhigh-density impulse noiseimpulse noise removal |
| spellingShingle | Trung-Hieu Le Po-Hsiung Lin Shih-Chia Huang LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network IEEE Open Journal of the Computer Society LD-Net lightweight denoising model high-density impulse noise impulse noise removal |
| title | LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network |
| title_full | LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network |
| title_fullStr | LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network |
| title_full_unstemmed | LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network |
| title_short | LD-Net: An Efficient Lightweight Denoising Model Based on Convolutional Neural Network |
| title_sort | ld net an efficient lightweight denoising model based on convolutional neural network |
| topic | LD-Net lightweight denoising model high-density impulse noise impulse noise removal |
| url | https://ieeexplore.ieee.org/document/9152117/ |
| work_keys_str_mv | AT trunghieule ldnetanefficientlightweightdenoisingmodelbasedonconvolutionalneuralnetwork AT pohsiunglin ldnetanefficientlightweightdenoisingmodelbasedonconvolutionalneuralnetwork AT shihchiahuang ldnetanefficientlightweightdenoisingmodelbasedonconvolutionalneuralnetwork |