Structurally and semantically coherent deep image inpainting
This thesis is submitted in partial fulfilment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2018.
| Main Authors: | , , |
|---|---|
| Outros autores: | |
| Formato: | Thesis |
| Idioma: | English |
| Publicado: |
BRAC University
2019
|
| Subjects: | |
| Acceso en liña: | http://hdl.handle.net/10361/11445 |
| id |
10361-11445 |
|---|---|
| record_format |
dspace |
| spelling |
10361-114452022-01-26T10:10:31Z Structurally and semantically coherent deep image inpainting Sami, Mirza Tanzim Khan, Ehsanul Amin Rhidita, Ishrat Naiyer Uddin, Jia Department of Computer Science and Engineering, BRAC University Image inpainting DCGAN Deep learning Differential equations, Partial. Image restoration. This thesis is submitted in partial fulfilment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2018. Cataloged from PDF version of thesis. Includes bibliographical references (pages 28-30). This thesis proposes an augmented method for image completion, particularly for images of human faces by leveraging on deep learning based inpainting techniques. Face completion generally tend to be a daunting task because of the relatively low uniformity of a face attributed to structures like eyes, nose, etc. Here, understanding the top level context is paramount for proper semantic completion. Our method improves upon existing inpainting techniques that reduces context difference by locating the closest encoding of the damaged image in the latent space of a pretrained deep generator. However, these existing methods fail to consider key facial structures (eyes, nose, jawline, etc) and their respective locations. We mitigate this by introducing a face landmark detector and a corresponding landmark loss. We add this landmark loss to the construction loss between the damaged and generated image and the adversarial loss of the generative model. After several experimentation, we concluded that the added landmark loss attributes to better understanding of top level context and hence more visually appealing inpainted images. Mirza Tanzim Sami Ehsanul Amin Khan Ishrat Naiyer Rhidita B. Computer Science and Engineering 2019-02-24T05:46:23Z 2019-02-24T05:46:23Z 2018 2018-12 Thesis ID 17141019 ID 14101118 ID 14310008 http://hdl.handle.net/10361/11445 en BRAC University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. 30 pages application/pdf BRAC University |
| institution |
Brac University |
| collection |
Institutional Repository |
| language |
English |
| topic |
Image inpainting DCGAN Deep learning Differential equations, Partial. Image restoration. |
| spellingShingle |
Image inpainting DCGAN Deep learning Differential equations, Partial. Image restoration. Sami, Mirza Tanzim Khan, Ehsanul Amin Rhidita, Ishrat Naiyer Structurally and semantically coherent deep image inpainting |
| description |
This thesis is submitted in partial fulfilment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2018. |
| author2 |
Uddin, Jia |
| author_facet |
Uddin, Jia Sami, Mirza Tanzim Khan, Ehsanul Amin Rhidita, Ishrat Naiyer |
| format |
Thesis |
| author |
Sami, Mirza Tanzim Khan, Ehsanul Amin Rhidita, Ishrat Naiyer |
| author_sort |
Sami, Mirza Tanzim |
| title |
Structurally and semantically coherent deep image inpainting |
| title_short |
Structurally and semantically coherent deep image inpainting |
| title_full |
Structurally and semantically coherent deep image inpainting |
| title_fullStr |
Structurally and semantically coherent deep image inpainting |
| title_full_unstemmed |
Structurally and semantically coherent deep image inpainting |
| title_sort |
structurally and semantically coherent deep image inpainting |
| publisher |
BRAC University |
| publishDate |
2019 |
| url |
http://hdl.handle.net/10361/11445 |
| work_keys_str_mv |
AT samimirzatanzim structurallyandsemanticallycoherentdeepimageinpainting AT khanehsanulamin structurallyandsemanticallycoherentdeepimageinpainting AT rhiditaishratnaiyer structurallyandsemanticallycoherentdeepimageinpainting |
| _version_ |
1814307784779366400 |