Object discovery via layer disposal
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017.
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | eng |
| Published: |
Massachusetts Institute of Technology
2018
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/1721.1/119538 |
| _version_ | 1811084886986981376 |
|---|---|
| author | Oktay, Deniz, M. Eng. Massachusetts Institute of Technology |
| author2 | Antonio Torralba. |
| author_facet | Antonio Torralba. Oktay, Deniz, M. Eng. Massachusetts Institute of Technology |
| author_sort | Oktay, Deniz, M. Eng. Massachusetts Institute of Technology |
| collection | MIT |
| description | Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017. |
| first_indexed | 2024-09-23T12:59:04Z |
| format | Thesis |
| id | mit-1721.1/119538 |
| institution | Massachusetts Institute of Technology |
| language | eng |
| last_indexed | 2024-09-23T12:59:04Z |
| publishDate | 2018 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1195382019-04-10T15:22:29Z Object discovery via layer disposal Oktay, Deniz, M. Eng. Massachusetts Institute of Technology Antonio Torralba. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Electrical Engineering and Computer Science. Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017. This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. Cataloged from student-submitted PDF version of thesis. Includes bibliographical references (pages 51-54). A key limitation of semantic image segmentation approaches is that they require large amounts of densely labeled training data. In this thesis, we introduce a method to learn to segment images with unlabeled data. The intuition behind the approach is that removing objects from images will yield natural images, however removing random patches will yield unnatural images. We capitalize on this signal to develop an auto-encoder that decomposes an image into layers, and when all layers are combined, it reconstructs the input image. However, when a layer is removed, the model learns to produce a different image that still looks natural to an adversary, which is possible by removing objects. Experiments and visualizations suggest that this model automatically learns to segment objects in images better than baselines. Some parts of this thesis represent joint work with Dr. Carl Vondrick and Professor Antonio Torralba. by Deniz Oktay. M. Eng. 2018-12-11T20:39:16Z 2018-12-11T20:39:16Z 2017 2017 Thesis http://hdl.handle.net/1721.1/119538 1076269771 eng MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission. http://dspace.mit.edu/handle/1721.1/7582 54 pages application/pdf Massachusetts Institute of Technology |
| spellingShingle | Electrical Engineering and Computer Science. Oktay, Deniz, M. Eng. Massachusetts Institute of Technology Object discovery via layer disposal |
| title | Object discovery via layer disposal |
| title_full | Object discovery via layer disposal |
| title_fullStr | Object discovery via layer disposal |
| title_full_unstemmed | Object discovery via layer disposal |
| title_short | Object discovery via layer disposal |
| title_sort | object discovery via layer disposal |
| topic | Electrical Engineering and Computer Science. |
| url | http://hdl.handle.net/1721.1/119538 |
| work_keys_str_mv | AT oktaydenizmengmassachusettsinstituteoftechnology objectdiscoveryvialayerdisposal |