Content Based Image Retrieval Using HDMR Constant Term Based Clustering

Karcılı A., Tunga B.

2nd International Conference on Mathematics and its Applications in Science and Engineering, ICMASE 2021, Salamanca, Spain, 1 - 02 July 2021, vol.384, pp.35-45 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 384
  • Doi Number: 10.1007/978-3-030-96401-6_3
  • City: Salamanca
  • Country: Spain
  • Page Numbers: pp.35-45
  • Keywords: Clustering, HDMR, Image decomposition, Image retrieval, k-means
  • Istanbul Technical University Affiliated: Yes

Abstract

© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.The studies related with the content-based image retrieval (CBIR) has increased because of both necessity for efficient image retrieval and the limitations in large-scale systems. Efficient image retrieval refers to finding accurate image from the database with high speed. This paper presents a new efficient image retrieval method using High Dimensional Model Representation (HDMR). The method has two main steps, clustering and retrieval. In clustering part, we use k-means method on HDMR constant term while in the subsequent part, we retrieve the most similar images to a given query image from a relevant cluster. We experiment the efficiency and effectiveness of the new algorithm on Columbia Object Image Library (COIL-100) and get conspicuous results. These results are tabulated in the paper.