Object Recognition using Multiresolution Transforms

International Journal of Research and Scientific Innovation (IJRSI) | Volume V, Issue V, May 2018 | ISSN 2321–2705

Object Recognition using Multiresolution Transforms

Ahila Priyadharshini. R

Dept. of Electronics and Communication Engineering, Mepco Schlenk Engineering College, Sivakasi – 626 005, Tamil Nadu, India

Abstract: Recognizing the objects in images is a challenging task due to the presence of occlusion, clutter, variation in shape, scale, color, illumination, position and size of objects in an image. In this paper, the potential efficiency of mutiresolution transforms such as Ridglet transform and Log Gabor transform for the Object Recognition task is investigated. To classify objects from images,local features such as patches are extracted over the interest points detected from the original image using Wavelet based interest point detector. Then Ridgelet features and Log Gabor features are computed for each and every patch. Then these features are trained, tested and classified using SVM classifier. The experimental evaluation of proposed method is done using the Graz01 database.

Key Words: Object Recognition, Salient Point, Patch, Log Gabor features, Ridgelet Features.

I. INTRODUCTION

Object recognition is of greater task in computer vision. It is the task of finding an object in image or video sequence. In case of humans, recognition task is much easier, that is they recognize millions of objects with little effort even when the objects look different in different circumstances. They could also recognize objects that are partially obstructed from the view. It is still a challenging task for computer systems to recognize objects that show different appearances in different surroundings [1].

Global features describe image as a whole and are less successful in recognition. Salient points are the points which maximize the discrimination between the objects. Salient point detection plays an important role in content based image retrieval in order to represent the local properties of the image. Since classic corner detectors cannot support natural images, detector based on wavelet transform represents global variations and local ones to detect the salient points [2, 3]. Schmid and Mohr (1997) proposed Local gray invariants for Image Retrieval, where local gray invariants are automatically extracted over the detected salient points [4]. Weber et al (2000) proposed the computation of K-means clustering algorithm at Forstner points for object recognition [5].

Read More