Relevance Feedback and Term Weighting Schemes for Content-Based Image Retrieval
VISUAL '99 Proceedings of the Third International Conference on Visual Information and Information Systems
Similarity learning via dissimilarity space in CBIR
MIR '06 Proceedings of the 8th ACM international workshop on Multimedia information retrieval
A similarity-based neural network for facial expression analysis
Pattern Recognition Letters
Annotating historical archives of images
Proceedings of the 8th ACM/IEEE-CS joint conference on Digital libraries
On automated image choice for secure and usable graphical passwords
Proceedings of the 28th Annual Computer Security Applications Conference
Hi-index | 0.00 |
In this paper we employ human judgments of image similarityto improve the organization of an image database.We first derive a statistic, \kappaB which measures the agreementbetween two partitionings of an image set. \kappaB is usedto assess agreement both amongst and between human andmachine partitionings. This provides a rigorous means ofchoosing between competing image database organizationsystems, and of assessing the performance of such systemswith respect to human judgments.Human partitionings of an image set are used to definean similarity value based on the frequency with which imagesare judged to be similar. When this measure is usedto partition an image set using a clustering technique, theresultant partitioning agrees better with human partitioningsthan any of the feature-space-based techniques investigated.Finally, we investigate the use multilayer perceptronsand a Distance Learning Network to learn a mapping fromfeature space to this perceptual similarity space. The DistanceLearning Network is shown to learn a mapping whichresults in partitionings in excellent agreement with thoseproduced by human subjects.