An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network
The process of iris recognition can result in a decline in recognition performance when the resolution of the iris images is insufficient. In this study, a super-resolution model for iris images, namely SwinGIris, which combines the Swin Transformer and the Generative Adversarial Network (GAN), is i...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-02-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/17/3/92 |
| _version_ | 1827307379596722176 |
|---|---|
| author | Hexin Lu Xiaodong Zhu Jingwei Cui Haifeng Jiang |
| author_facet | Hexin Lu Xiaodong Zhu Jingwei Cui Haifeng Jiang |
| author_sort | Hexin Lu |
| collection | DOAJ |
| description | The process of iris recognition can result in a decline in recognition performance when the resolution of the iris images is insufficient. In this study, a super-resolution model for iris images, namely SwinGIris, which combines the Swin Transformer and the Generative Adversarial Network (GAN), is introduced. SwinGIris performs quadruple super-resolution reconstruction for low-resolution iris images, aiming to improve the resolution of iris images and thereby improving the recognition accuracy of iris recognition systems. The model utilizes residual Swin Transformer blocks to extract depth global features, and the progressive upsampling method along with sub-pixel convolution is conducive to focusing on the high-frequency iris information in the presence of more non-iris information. In order to preserve high-frequency details, the discriminator employs a VGG-style relative classifier to guide the generator in generating super-resolution images. In experimental section, we enhance low-resolution (56 × 56) iris images to high-resolution (224 × 224) iris images. Experimental results indicate that the SwinGIris model achieves satisfactory outcomes in restoring low-resolution iris image textures while preserving identity information. |
| first_indexed | 2024-04-24T18:38:03Z |
| format | Article |
| id | doaj.art-be942f678dfa47098cb907905c1d6e28 |
| institution | Directory Open Access Journal |
| issn | 1999-4893 |
| language | English |
| last_indexed | 2024-04-24T18:38:03Z |
| publishDate | 2024-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj.art-be942f678dfa47098cb907905c1d6e282024-03-27T13:17:13ZengMDPI AGAlgorithms1999-48932024-02-011739210.3390/a17030092An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial NetworkHexin Lu0Xiaodong Zhu1Jingwei Cui2Haifeng Jiang3Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, ChinaKey Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, ChinaKey Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, ChinaKey Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, ChinaThe process of iris recognition can result in a decline in recognition performance when the resolution of the iris images is insufficient. In this study, a super-resolution model for iris images, namely SwinGIris, which combines the Swin Transformer and the Generative Adversarial Network (GAN), is introduced. SwinGIris performs quadruple super-resolution reconstruction for low-resolution iris images, aiming to improve the resolution of iris images and thereby improving the recognition accuracy of iris recognition systems. The model utilizes residual Swin Transformer blocks to extract depth global features, and the progressive upsampling method along with sub-pixel convolution is conducive to focusing on the high-frequency iris information in the presence of more non-iris information. In order to preserve high-frequency details, the discriminator employs a VGG-style relative classifier to guide the generator in generating super-resolution images. In experimental section, we enhance low-resolution (56 × 56) iris images to high-resolution (224 × 224) iris images. Experimental results indicate that the SwinGIris model achieves satisfactory outcomes in restoring low-resolution iris image textures while preserving identity information.https://www.mdpi.com/1999-4893/17/3/92iris recognitionimage reconstructionsuper-resolutiontransformerGenerative Adversarial Network (GAN) |
| spellingShingle | Hexin Lu Xiaodong Zhu Jingwei Cui Haifeng Jiang An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network Algorithms iris recognition image reconstruction super-resolution transformer Generative Adversarial Network (GAN) |
| title | An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network |
| title_full | An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network |
| title_fullStr | An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network |
| title_full_unstemmed | An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network |
| title_short | An Iris Image Super-Resolution Model Based on Swin Transformer and Generative Adversarial Network |
| title_sort | iris image super resolution model based on swin transformer and generative adversarial network |
| topic | iris recognition image reconstruction super-resolution transformer Generative Adversarial Network (GAN) |
| url | https://www.mdpi.com/1999-4893/17/3/92 |
| work_keys_str_mv | AT hexinlu anirisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT xiaodongzhu anirisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT jingweicui anirisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT haifengjiang anirisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT hexinlu irisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT xiaodongzhu irisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT jingweicui irisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork AT haifengjiang irisimagesuperresolutionmodelbasedonswintransformerandgenerativeadversarialnetwork |