DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lee, Eon-Ho | ko |
dc.contributor.author | Jeon, MyungHwan | ko |
dc.contributor.author | Jang, Hyesu | ko |
dc.contributor.author | Park, Byungjae | ko |
dc.contributor.author | Kim, Ayoung | ko |
dc.contributor.author | Lee, Sejin | ko |
dc.date.accessioned | 2023-08-25T06:00:49Z | - |
dc.date.available | 2023-08-25T06:00:49Z | - |
dc.date.created | 2023-07-06 | - |
dc.date.issued | 2020-09 | - |
dc.identifier.citation | 2020 IEEE/OES Autonomous Underwater Vehicles Symposium, AUV 2020 | - |
dc.identifier.issn | 1522-3167 | - |
dc.identifier.uri | http://hdl.handle.net/10203/311823 | - |
dc.description.abstract | This research is a study to check whether the synthesized experimental data in the laboratory without experimentation is useful as experimental data. In the paper, synthesized data is generated base on the pix2pix algorithm. The training data used in the FCN checks the performance of the FCN by adjusting the ratio between the synthetic data and the experimental data obtained through the experiment. Various cases are used to confirm that synthesized data can be replaced with data obtained through actual experimentation. | - |
dc.language | English | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.title | Study on the Training Effectiveness of Deep Learning with Synthesized Underwater Sonar Image Using Pix2Pix and FCN | - |
dc.type | Conference | - |
dc.identifier.wosid | 000896378600040 | - |
dc.identifier.scopusid | 2-s2.0-85098517237 | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | 2020 IEEE/OES Autonomous Underwater Vehicles Symposium, AUV 2020 | - |
dc.identifier.conferencecountry | CN | - |
dc.identifier.conferencelocation | St Johns | - |
dc.identifier.doi | 10.1109/AUV50043.2020.9267923 | - |
dc.contributor.localauthor | Kim, Ayoung | - |
dc.contributor.nonIdAuthor | Lee, Eon-Ho | - |
dc.contributor.nonIdAuthor | Park, Byungjae | - |
dc.contributor.nonIdAuthor | Lee, Sejin | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.