Please use this identifier to cite or link to this item:
http://artemis.cslab.ece.ntua.gr:8080/jspui/handle/123456789/18228
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Υψηλάντης, Νικόλαος-Αντώνιος | - |
dc.date.accessioned | 2022-02-22T10:27:19Z | - |
dc.date.available | 2022-02-22T10:27:19Z | - |
dc.date.issued | 2022-02-15 | - |
dc.identifier.uri | http://artemis.cslab.ece.ntua.gr:8080/jspui/handle/123456789/18228 | - |
dc.description.abstract | In this thesis, the creation of a new dataset and benchmark for large-scale instance-level recognition (ILR) in the domain of artworks is addressed. The proposed dataset, called the Met dataset, exhibits a number of different challenges such as large inter-class similarity, long-tail distribution, and many classes. It relies on the open access collection of The Met museum to form a large training set of about 224k classes, where each class corresponds to a museum exhibit with photos taken under studio conditions. The evaluation set is primarily composed of photos taken by museum guests depicting exhibits, which introduces a distribution shift between training and testing. It is additionally composed of a set of images not related to Met exhibits making the task resemble an out-of-distribution detection problem. The proposed benchmark follows the paradigm of other recent datasets for ILR on different domains to encourage research on domain independent approaches. In order to offer a testbed for future comparisons, a number of suitable approaches are evaluated. Self-supervised and supervised contrastive learning are effectively combined to train CNNs that produce the image representation used in combination with a non-parametric classifier, showing a promising direction for ILR. The dataset webpage (also contains reference code) is: http://cmp.felk.cvut.cz/met/. | en_US |
dc.language | en | en_US |
dc.subject | instance-level recognition | en_US |
dc.subject | artwork recognition | en_US |
dc.subject | large-scale recognition dataset | en_US |
dc.subject | knn classification | en_US |
dc.title | Instance-level recognition for artworks | en_US |
dc.description.pages | 85 | en_US |
dc.contributor.supervisor | Κόλλιας Στέφανος | en_US |
dc.department | Τομέας Τεχνολογίας Πληροφορικής και Υπολογιστών | en_US |
Appears in Collections: | Διπλωματικές Εργασίες - Theses |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Thesis_Nikolaos-Antonios_Ypsilantis.pdf | 45.12 MB | Adobe PDF | View/Open |
Items in Artemis are protected by copyright, with all rights reserved, unless otherwise indicated.