Neural network-based shape retrieval using moment invariants and Zernike moments.

Date of Award


Degree Type


Degree Name



Computer Science


Computer Science.




Shape is one of the fundamental image features for use in Content-Based Image Retrieval (CBIR). Compared with other visual features such as color and texture, it is extremely powerful and provides capability for object recognition and similarity-based image retrieval. In this thesis, we propose a Neural Network-Based Shape Retrieval System using Moment Invariants and Zernike Moments. Moment Invariants and Zernike Moments are two region-based shape representation schemes and are derived from the shape in an image and serve as image features. k means clustering is used to group similar images in an image collection into k clusters whereas Neural Network is used to facilitate retrieval against a given query image. Neural Network is trained by the clustering result on all of the images in the collection using back-propagation algorithm. In this scheme, Neural Network serves as a classifier such that moments are inputs to the Neural Network and the output is one of the k classes that have the largest similarities to the query image. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis2005 .C444. Source: Masters Abstracts International, Volume: 44-03, page: 1396. Thesis (M.Sc.)--University of Windsor (Canada), 2005.