Object recognition using Gabor wavelet features with various classification techniques

Abstract

Object recognition is a process of detection and recognition of certain classes of objects like chairs, guitars, buildings etc., from an image or video sequence. Various researches have been done till now for extraction of object’s features from real-world images using various approaches varying from appearance based approaches like PCA, LDA, ICA, moment invariant, shape context, SIFT etc. or model based approaches. We used a well known model based approach ‘‘Gabor wavelet transform’’ to extract the features. Gabor wavelets exhibit desirable characteristics of spatial locality and orientation selectivity. It has several advantages against robustness, illumination, multi-resolution, and multi-orientation. Classifications of objects are important areas in a variety of fields, such as pattern recognition, artificial intelligence and vision analysis. So using the Gabor wavelet features classification is done by various well known classifiers like KNN, Neural Network (NN), SVM, and Naive Bayes classifiers. So results are compared using these classifiers as well we have given various pros and cons for these methods.