DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kweon, In So | - |
dc.contributor.advisor | 권인소 | - |
dc.contributor.author | Karjauv, Adil | - |
dc.date.accessioned | 2022-04-21T19:32:33Z | - |
dc.date.available | 2022-04-21T19:32:33Z | - |
dc.date.issued | 2021 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=963416&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/295520 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2021.8,[iv, 22 p. :] | - |
dc.description.abstract | Data hiding is one widely used approach for proving ownership through blind watermarking. Deep learning has been widely used in data hiding and inserting an attack simulation layer (ASL) after the watermarked image has been widely recognized as the most effective approach for improving the pipeline robustness against distortions. Despite its wide usage due to simplicity, the gain of enhanced robustness is usually interpreted through the lens of augmentation while our work explores this gain from a new perspective by disentangling the forward and backward pass of such ASL. We find that the main influential component is the forward pass instead of the backward pass. This observation motivates us to use forward ASL to make the pipeline compatible with non-differentiable and/or black-box distortion, such as lossy (JPEG) compression and photoshop effects. Extensive experiments demonstrate the efficacy of our approach despite being simple. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Robust Deep Hiding▼aLossy Compression▼aNon-differentiable▼aBlind Watermarking | - |
dc.subject | 강인한 심층 은닉▼a손실 압축▼a미분불가능한 블라인드 워터마킹 | - |
dc.title | Towards robust deep hiding under non-differentiable distortions for practical blind watermarking | - |
dc.title.alternative | 실속형 블라인드 워터마킹을 위한 미분불가능한 왜곡에 의한 강력한 강인한 심층 은닉에 대한 연구 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
dc.contributor.alternativeauthor | 카르자우프 아딜 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.