An Efficient Image Retrieval Model with Convolutional Neural Network based Text/Image Identification for Copyright Violation Detection Telif Hakki Ihlali Tespiti için Evrişimsel Sinir Aǧi Tabanli Metin/Görüntü Tanimlamali Verimli Bir Görüntü Geri Getirme Modeli

Ozden H., Tavli B., DEMİRCİ M.

30th Signal Processing and Communications Applications Conference, SIU 2022, Safranbolu, Turkey, 15 - 18 May 2022, (Full Text) identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu55565.2022.9864741
  • City: Safranbolu
  • Country: Turkey
  • Keywords: Bag of Visual Words, Convolutional Neural Network, Image Retrieval, SIFT
  • Ankara Yıldırım Beyazıt University Affiliated: Yes

Abstract

© 2022 IEEE.One of the most important problems faced by broadcasters is the unauthorized use of their images by third parties or organizations in a large-scale database, which contains hundreds of thousands of images. For this reason, it is important to perform an efficient and effective image retrieval, whose objective is to find the most similar images to a given test image. In addition, test images often contain text, and the presence of the text together with the visual part complicates the search process. In this paper, we present an image retrieval framework based on a bag of visual words, which has been shown to be effective in the literature. A convolutional neural network model is used to parse the text in the images. Experiments demonstrate the efficacy of this model in a large database.