Pre-classification module for an all-season image retrieval system

Hong FU, Zheru CHI, Dagan FENG, Weibao ZOU, King Chuen LO, Xiaoyu ZHAO

Research output: Chapter in Book/Report/Conference proceedingChapters

2 Citations (Scopus)

Abstract

From the study of attention-driven image interpretation and retrieval, we have found that an attention-driven strategy is able to extract important objects from an image and then focus the attentive objects while retrieving images. However, besides the images with distinct objects, there are images which do not show distinct objects. In this paper, the classification of "attentive" and "non-attentive" image is proposed to be a pre-process module in an all-season image retrieval system which can tackle both kinds of images. In this pre-classification module, an image is represented by an adaptive tree structure with each node carrying normalized features that characterize the object/region with visual contrasts and spatial information. Then a neural network is trained to classify an image as an " attentive" or "non-attentive" category by using the Back Propagation Through Structure (BPTS) algorithm. Experimental results indicate the reliability and feasibility of the pre-classification module, which encourages us to conduct further investigations on the all-season image retrieval system. Copyright © 2007 IEEE.
Original languageEnglish
Title of host publicationThe 2007 International Joint Conference on Neural Networks: IJCNN 2007 conference proceedings
Place of PublicationPiscataway, NJ
PublisherIEEE
Pages2642-2646
ISBN (Electronic)9781424413805
ISBN (Print)9781424413799
DOIs
Publication statusPublished - 2007

Citation

Fu, H., Chi, Z., Feng, D., Zou, W., Lo, K. C., & Zhao, X. (2007). Pre-classification module for an all-season image retrieval system. In The 2007 International Joint Conference on Neural Networks: IJCNN 2007 conference proceedings (pp. 2642-2646). Piscataway, NJ: IEEE.

Fingerprint

Dive into the research topics of 'Pre-classification module for an all-season image retrieval system'. Together they form a unique fingerprint.