Analyzing deep features for trademark retrieval Marka Erişimi İcin Derin Özniteliklerin İncelenmesi

Date

2017-05-18

Author

Aker, Cemal
Tursun, Osman
Kalkan, Sinan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

161
views

0
downloads

The rapid rise in the amount of trademark applications and trademark infringements has led the trademark retrieval (TR) to become an important and formidable task to solve. Existing studies based on hand-crafted features show unsatisfying performance. Taking the popularization and increasing success of the deep learning methods into consideration, in this work, many well-known Convolutional Neural Network (CNN) models are applied to the TR problem. Models are tested using a large scale trademark dataset in contrast with the previously proposed solutions, and their failure points are discussed in this study. For the problems that can be encountered, solutions such as tine-tuning, distance metric learning, using CNN features locally, and making them invariant to aspect ratio of the trademark are suggested.

Subject Keywords

Deep learning, Trademark retrieval

URI

https://hdl.handle.net/11511/35921

DOI

https://doi.org/10.1109/siu.2017.7960426

Collections

Department of Computer Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

METU Dataset: A Big Dataset for Benchmarking Trademark Retrieval Tursun, Osman; Kalkan, Sinan (2015-05-22) Trademark retrieval (TR) is the problem of retrieving similar trademarks (logos) for a query, and the main aim is to detect copyright infringements in trademarks. Since there are millions of companies worldwide, automatically retrieving similar trademarks has become an important problem, and currently, checking trademark infringements is mostly performed manually by humans. However, although there have been many attempts for automated TR, as also acknowledged in the community, the problem is largely unsolve...
Investigating open design using jugaad as a cultural probe Hasan, Aleena; Pedgley, Owaın Francıs; Department of Industrial Design (2021-12-01) Design democratization has been a growing movement around the world, blurring the boundary between the roles of users and designers and encouraging design-after-design that makes the process of designing more inclusive. Open Design is one such approach that encourages no limits in terms of the kind, time, or space of the contribution. While a very primary attribute of this approach is inclusivity, there is a lack of exploration in terms of actual application of the process in a tangible product design proce...
Proposal for a non-dimensional parametric interface design in architecture : a biomimetic approach Arslan Selçuk, Semra; Sorguç, Arzu; Department of Building Science in Architecture (2009) Biomimesis, the imitation of animate and inanimate forms in nature to inspire new designs, is term introduced in the 20th century. The concept that there exist models and solutions in nature that may improve and optimize the way mankind lives has been the subject of much discussion. Although biomimesis as a well-defined discipline is a relatively recent concept, modeling nature is as old as mankind itself and can be seen in many different forms in all aspects of life. In the field of architecture there have...
TestBATN - a scenario based test platform for conformance and interoperability testing Namlı, Tuncay; Doğaç, Asuman; Department of Computer Engineering (2011) Today, interoperability is the major challenge for e-Business and e-Government domains. The fundamental solution is the standardization in different levels of business-to-business interactions. However publishing standards alone are not enough to assure interoperability between products of different vendors. In this respect, testing and certification activities are very important to promote standard adoption, validate conformance and interoperability of the products and maintain correct information exchange...
Wasserstein generative adversarial active learning for anomaly detection with gradient penalty Duran, Hasan Ali; Ertekin Bolelli, Şeyda; Department of Computer Engineering (2021-9) Anomaly detection has become a very important topic with the advancing machine learning techniques and is used in many different application areas. In this study, we approach differently than the anomaly detection methods performed on standard generative models and describe anomaly detection as a binary classification problem. However, in order to train a highly accurate classifier model, the number of anomaly data in data-sets is very limited, and with synthetic data produced using generative models, it ca...

Citation Formats

C. Aker, O. Tursun, and S. Kalkan, “Analyzing deep features for trademark retrieval Marka Erişimi İcin Derin Özniteliklerin İncelenmesi,” 2017, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/35921.