Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value
Image-to-image conversion tasks are more accurate and sophisticated than ever thanks to advances in deep learning. However, since typical deep learning models are trained to perform only one task, multiple trained models are required to perform each task even if they are related to each other. For e...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9481244/ |
| _version_ | 1819094809542918144 |
|---|---|
| author | Seokjun Kang Seiichi Uchida Brian Kenji Iwana |
| author_facet | Seokjun Kang Seiichi Uchida Brian Kenji Iwana |
| author_sort | Seokjun Kang |
| collection | DOAJ |
| description | Image-to-image conversion tasks are more accurate and sophisticated than ever thanks to advances in deep learning. However, since typical deep learning models are trained to perform only one task, multiple trained models are required to perform each task even if they are related to each other. For example, the popular image-to-image convolutional neural network, U-Net, is normally trained for a single task. Based on U-Net, this study proposes a model that outputs variable results using only one trained model. The proposed method produces a continuously changing output by setting an external parameter. We confirm the robustness of our proposed model by evaluating it on binarization and background blurring. According to these evaluations, we confirmed that the proposed model can generate well-predicted outputs by using un-trained tuning parameters as well as the outputs by using trained tuning parameters. Furthermore, the proposed model can generate extrapolated outputs outside the learning range. |
| first_indexed | 2024-12-21T23:33:18Z |
| format | Article |
| id | doaj.art-9fabea66755047678da871ad75e82ae7 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-21T23:33:18Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-9fabea66755047678da871ad75e82ae72022-12-21T18:46:27ZengIEEEIEEE Access2169-35362021-01-01910327910329010.1109/ACCESS.2021.30965309481244Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar ValueSeokjun Kang0https://orcid.org/0000-0002-3316-0532Seiichi Uchida1https://orcid.org/0000-0001-8592-7566Brian Kenji Iwana2https://orcid.org/0000-0002-5146-6818Department of Advanced Information Technology, Kyushu University, Fukuoka, JapanDepartment of Advanced Information Technology, Kyushu University, Fukuoka, JapanDepartment of Advanced Information Technology, Kyushu University, Fukuoka, JapanImage-to-image conversion tasks are more accurate and sophisticated than ever thanks to advances in deep learning. However, since typical deep learning models are trained to perform only one task, multiple trained models are required to perform each task even if they are related to each other. For example, the popular image-to-image convolutional neural network, U-Net, is normally trained for a single task. Based on U-Net, this study proposes a model that outputs variable results using only one trained model. The proposed method produces a continuously changing output by setting an external parameter. We confirm the robustness of our proposed model by evaluating it on binarization and background blurring. According to these evaluations, we confirmed that the proposed model can generate well-predicted outputs by using un-trained tuning parameters as well as the outputs by using trained tuning parameters. Furthermore, the proposed model can generate extrapolated outputs outside the learning range.https://ieeexplore.ieee.org/document/9481244/Image-to-image conversionmultiple tasksU-Netimage binarizationbackground blur |
| spellingShingle | Seokjun Kang Seiichi Uchida Brian Kenji Iwana Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value IEEE Access Image-to-image conversion multiple tasks U-Net image binarization background blur |
| title | Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value |
| title_full | Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value |
| title_fullStr | Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value |
| title_full_unstemmed | Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value |
| title_short | Tunable U-Net: Controlling Image-to-Image Outputs Using a Tunable Scalar Value |
| title_sort | tunable u net controlling image to image outputs using a tunable scalar value |
| topic | Image-to-image conversion multiple tasks U-Net image binarization background blur |
| url | https://ieeexplore.ieee.org/document/9481244/ |
| work_keys_str_mv | AT seokjunkang tunableunetcontrollingimagetoimageoutputsusingatunablescalarvalue AT seiichiuchida tunableunetcontrollingimagetoimageoutputsusingatunablescalarvalue AT briankenjiiwana tunableunetcontrollingimagetoimageoutputsusingatunablescalarvalue |