Hybrid Approach to Content-Based Image Retrieval Using Modified Multi-Scale LBP and Color Features
- PDF / 1,621,339 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 58 Downloads / 232 Views
ORIGINAL RESEARCH
Hybrid Approach to Content‑Based Image Retrieval Using Modified Multi‑Scale LBP and Color Features Sagar Chavda1 · Mahesh Goyani1 Received: 25 May 2020 / Accepted: 5 September 2020 © Springer Nature Singapore Pte Ltd 2020
Abstract The objective of the content-based image retrieval (CBIR) system is to retrieve the visually identical images from the database efficiently and effectively. It is a broad research realm with the availability of numerous applications. Performance dependence of CBIR focuses on the extraction, reduction, and selection of the features along with the practice of classification technique. In this work, we have proposed the hybrid approach of two different feature descriptors: global color histogram and multi-scale local binary pattern (MS-LBP); furthermore, the use of PCA for dimension reduction and LDA for the selection of features. The proposed method is evaluated concerning various benchmark datasets, viz., Corel-1k, Corel-5k, Corel-10k, and Ghim-10k together with result comparison based on the precision and recall values at different thresholds. The classification purposes are satisfied with Euclidean and City Block distance. The performance study of the proposed work displays it as outperformer than the identified state-of-the-art literature. Keywords CBIR · Color histogram · MS-LBP · PCA · LDA
Introduction Content-based image retrieval (CBIR) is a solution to the image retrieval problem using the contents of the images. Image retrieval is defined as the retrieval of semantically relevant images from a database of images. The requirement of retrieving similar images from a large database is the need of the hour; to surpass this hiccup, there should be some explanation and quick fix. Images can be retrieved based on the metadata or content of the query image. Assigning naming in traditional image retrieval systems is a clumsy task. As time goes, retrieving, processing, searching, browsing, and managing of images became hard. To solve the image retrieval problem, images are retrieved based on high-level or low-level features or sometimes a combination of both. Retrieving the images based on a query from the large database is not an easy task when you are adopting a traditional approach like text-based image retrieval (TBIR) where the
* Sagar Chavda [email protected] Mahesh Goyani [email protected] 1
Gujarat Technological University, Ahmedabad, India
search is based on automatic or manual annotation of images. The TBIR retrieves the images from the database based on the metadata of the query image. Therefore, the problem of annotation of the images and laboring cost arises. The most known problems with the TBIR are annotation of the images, human perception, and deeper needs of the queries for searching images [1]. To overcome the above disadvantages in TBIR, content-based approach was introduced in the early 1980s. CBIR is the use of computer vision to the image retrieval difficulty that is the crisis of searching for digital images in huge databases. In
Data Loading...
