Title: Study of Zernike Moments and Complex Zernike Moments for Improving
Retrieval Accuracy in Shape Based Image Retrieval
Publisher: Guru Nanak Publications
Series: Volume 2 Issue 2
Authors: Anjali Goyal
Retrieval of images on shape basis is required in
case of logo, trademark, clip art search. Shape is one of the
best criteria to retrieve the images from a large database
when color and texture of data comes at secondary place.
To extract the features, different techniques are used.
Zernike Moments (ZMs) has been most widely used
technique in extracting the region based shape features of
an image. But retrieval of an accurate image is an issue. In
most of the research, emphasis is given only on magnitude
of ZMs. Phase component has been ignored. In the paper,
complex ZMs based features are compared with the ZMs
based features. A comparative study has been made to see
the retrieval performance. Angle and magnitude based
Distance has been used to see the similarity of images.
Shape database of 50 images of 10 different classes is used
in the study. Rotational database is generated using these
images. It is found that Complex ZMs can capture better
features as compared to magnitude based ZMs. The
implementation was done in VC++ environment.
Magnitude, Phase, Shape based image retrieval (SBIR), Zernike moments.