MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal
Since machine learning techniques for raindrop removal have not been capable of completely removing raindrops and have failed to take into account the constraints of edge devices with limited resources, a novel software-hardware co-designed method with a memristor for raindrop removal, named memrist...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-01-01
|
| Series: | Micromachines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-666X/15/2/217 |
| _version_ | 1797297492153335808 |
|---|---|
| author | Qiuyue Chai Yue Liu |
| author_facet | Qiuyue Chai Yue Liu |
| author_sort | Qiuyue Chai |
| collection | DOAJ |
| description | Since machine learning techniques for raindrop removal have not been capable of completely removing raindrops and have failed to take into account the constraints of edge devices with limited resources, a novel software-hardware co-designed method with a memristor for raindrop removal, named memristive attention recurrent residual generative adversarial network (MARR-GAN), is introduced in this research. A novel raindrop-removal network is specifically designed based on attention gate connections and recurrent residual convolutional blocks. By replacing the basic convolution unit with recurrent residual convolution unit, improved capturing of the changes in raindrop appearance over time is achieved, while preserving the position and shape information in the image. Additionally, an attention gate is utilized instead of the original skip connection to enhance the overall structural understanding and local detail preservation, facilitating a more comprehensive removal of raindrops across various areas of the image. Furthermore, a hardware implementation scheme for MARR-GAN is presented in this paper, where deep learning algorithms are seamlessly integrated with neuro inspired computing chips, utilizing memristor crossbar arrays for accelerated real-time image-data processing. Compelling evidence of the efficacy and superiority of MARR-GAN in raindrop removal and image restoration is provided by the results of the empirical study. |
| first_indexed | 2024-03-07T22:21:09Z |
| format | Article |
| id | doaj.art-a6de8efc13e94c5aab6bd0116ba98d97 |
| institution | Directory Open Access Journal |
| issn | 2072-666X |
| language | English |
| last_indexed | 2024-03-07T22:21:09Z |
| publishDate | 2024-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Micromachines |
| spelling | doaj.art-a6de8efc13e94c5aab6bd0116ba98d972024-02-23T15:27:38ZengMDPI AGMicromachines2072-666X2024-01-0115221710.3390/mi15020217MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop RemovalQiuyue Chai0Yue Liu1School of Electrical and Electronic Engineering, Changchun University of Technology, Changchun 130012, ChinaSchool of Electrical and Electronic Engineering, Changchun University of Technology, Changchun 130012, ChinaSince machine learning techniques for raindrop removal have not been capable of completely removing raindrops and have failed to take into account the constraints of edge devices with limited resources, a novel software-hardware co-designed method with a memristor for raindrop removal, named memristive attention recurrent residual generative adversarial network (MARR-GAN), is introduced in this research. A novel raindrop-removal network is specifically designed based on attention gate connections and recurrent residual convolutional blocks. By replacing the basic convolution unit with recurrent residual convolution unit, improved capturing of the changes in raindrop appearance over time is achieved, while preserving the position and shape information in the image. Additionally, an attention gate is utilized instead of the original skip connection to enhance the overall structural understanding and local detail preservation, facilitating a more comprehensive removal of raindrops across various areas of the image. Furthermore, a hardware implementation scheme for MARR-GAN is presented in this paper, where deep learning algorithms are seamlessly integrated with neuro inspired computing chips, utilizing memristor crossbar arrays for accelerated real-time image-data processing. Compelling evidence of the efficacy and superiority of MARR-GAN in raindrop removal and image restoration is provided by the results of the empirical study.https://www.mdpi.com/2072-666X/15/2/217raindrop removalrecurrent residual networkmemristorattention gateGAN |
| spellingShingle | Qiuyue Chai Yue Liu MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal Micromachines raindrop removal recurrent residual network memristor attention gate GAN |
| title | MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal |
| title_full | MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal |
| title_fullStr | MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal |
| title_full_unstemmed | MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal |
| title_short | MARR-GAN: Memristive Attention Recurrent Residual Generative Adversarial Network for Raindrop Removal |
| title_sort | marr gan memristive attention recurrent residual generative adversarial network for raindrop removal |
| topic | raindrop removal recurrent residual network memristor attention gate GAN |
| url | https://www.mdpi.com/2072-666X/15/2/217 |
| work_keys_str_mv | AT qiuyuechai marrganmemristiveattentionrecurrentresidualgenerativeadversarialnetworkforraindropremoval AT yueliu marrganmemristiveattentionrecurrentresidualgenerativeadversarialnetworkforraindropremoval |