Directional local ternary quantized extrema pattern: A new descriptor for biomedical image indexing and retrieval

Abstract

This paper proposes a new pattern descriptor called directional local ternary quantized extrema pattern (DLTerQEP) for biomedical image indexing and retrieval. The standard local binary patterns (LBPs) and local ternary patterns (LTPs) encode the gray scale relationship between the center pixel and its surrounding neighbors in two dimensional (2D) local region of an image whereas the proposed method encodes the spatial relation between any pair of neighbors in a local region along the given directions (i.e., 0°, 45°, 90° and 135°) for a given center pixel in an image. The novelty of the proposed method is it uses ternary patterns from horizontal–vertical–diagonal–antidiagonal (HVDA7) structure of directional local extrema values of an image to encode more spatial structure information which lead to better retrieval. DLTerQEP also provides a significant increase in discriminative power by allowing larger local pattern neighborhoods. The experiments have been carried out for proving the worth of proposed algorithm on three different types of benchmark biomedical databases; (i) computed tomography (CT) scanned lung image databases named as LIDC-IDRI-CT and VIA/I-ELCAP-CT, (ii) brain magnetic resonance imaging (MRI) database named as OASIS-MRI. The results demonstrate the superiority of the proposed method in terms of average retrieval precision (ARP) and average retrieval rate (ARR) over state-of-the-art feature extraction techniques like LBP, LTP and LQEP etc.