Indian Journal of Science and Technology
Year: 2017, Volume: 10, Issue: 18, Pages: 1-10
Harpreet Kaur1* and Vijay Dhir2
1Department of Computer Science and Engineering, Punjab Technical University, Kapurthala, Punjab - 144601, India; [email protected] 2Sant Baba Bhag Singh Insitute of Engineering & Technology, Jalandhar, Punjab - 144020, India
*Author for the correspondence:
Department of Computer Science and Engineering, Punjab Technical University, Kapurthala, Punjab - 144601, India; [email protected]
Objectives: This paper presents a new feature extractor which we named as Local Color Oppugnant Mesh Extrema Patterns (LCOMeEP) for retrieval of images. Methods/Statistical Analysis: The suggested method gathers the colortexture statistics among the RGB (Red, Green and Blue) and gray scale of the given image. The color-texture statistics is extracted based on mesh extrema which assembles the relationship among the neighbors using extrema. Findings: The proposed method is diverse from the Directional Local Extrema Patterns (DLEP) collect the directional information which is based on local extrema in an image. Whereas the method presented in this paper, (LCOMeEP) gathers the mesh extremas amid the RgG (red, gray, green), GgB (green, gray, blue) and BgR (blue, gray, red) spaces. The enactment of the presented technique is amended by assimilating the LCOMeEPs with histograms of HSV (hue, saturation, value). The enactment of the research work is estimated by simulating on standard data sets, Corel-5K and Corel-10K in context of recall, precision, Average Retrieval Rate (ARR) and Average Retrieval Precision (ARP). Application/Improvements: The outcome after inspection illustrates a substantial enhancement as related to the contemporary features for image retrieval.
Keywords: Image Retrieval, Local Binary Patterns (LBP), Local Extrema Patterns, Local Oppugnant Patterns, Pattern Recognition, Texture
Subscribe now for latest articles and news.