Title :
Design Image Retrieval Based on Nonsubsampled Contourlet Transform and Neural Networks
Author :
Li, Yi ; Liu, Guanzhong
Author_Institution :
Bus. Sch., Central South Univ., Changsha, China
Abstract :
In this paper, an image retrieval method based on nonsubsampled contourlet transform (NSCT) is proposed. Firstly, the image is decomposed into different scales and different directions subbands by the NSCT. Then, texture features and shape features of subbands are extracted. Finally, BP neural network is used to retrieve images through using extracted features. The experimental results over car accessory images demonstrate the effectiveness of the proposed method.
Keywords :
automobiles; automotive components; backpropagation; feature extraction; image retrieval; mechanical engineering computing; neural nets; transforms; BP neural network; NSCT; car accessory images; design image retrieval; nonsubsampled contourlet transform; shape feature extraction; texture feature extraction; Computed tomography; Feature extraction; Filter banks; Image retrieval; Matrix decomposition; Shape; Transforms;
Conference_Titel :
Management and Service Science (MASS), 2011 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-6579-8
DOI :
10.1109/ICMSS.2011.5998899
