Open AccessFeatures Extraction for Pollen Recognition Using Gabor Filters
Author(s) -
Dimitar Nikolov,
Diana Tsankova
Publication year - 2018
Publication title -
food science and applied biotechnology
Language(s) - English
Resource type - Journals
ISSN - 2603-3380
DOI - 10.30721/fsab2018.v1.i2.11
Subject(s) - artificial intelligence , pattern recognition (psychology) , computer science , gabor filter , normalization (sociology) , visualization , principal component analysis , pollen , computer vision , feature extraction , botany , biology , sociology , anthropology
The aim of the article is to investigate the features extraction from microscope images of pollens for a classification of honey on the base of its botanical origin. A filter-bank of Gabor filters (as a biologically inspired recognition system) is used to obtain features, which are then post-processed using normalization, down-sampling (by bicubic interpolation), and principal components analysis (PCA). PCA is used for reducing the features size and a proper visualization of the features extraction results. Microscope images from the European pollen database, including pollen images of linden, acacia, lavender, rapeseed, and thistle, are used to illustrate capabilities of the proposed features extraction approach. The performance of the proposed algorithm is evaluated by simulations in MATLAB environment.