Citation Context

...proximation of the color distribution. It gives a simple and compact representation of the color content of an artwork. This representation has been intensively used in image retrieval systems, e.g. [=-=Kushki et al. 2004-=-]. Instead of considering all colors appearing in an image, many papers tend to reduce the number of extracted colors. This type of approach is sustained by the fact that artists use a limited set of ...

Citation Context

...several local minima that are not trivial to minimize, especially with a deterministic optimizer. Another family of fusion approaches used in imagery are the so-called decision fusion approaches [30]–=-=[32]-=-. With these methods, a series of energy functions are first minimized before their outputs (their decisions) are merged. In this case, the energy functions are defined on different basic features and...

Citation Context

...ia domain. In order to address this issue, researchers from the computer vision community continuously propose new content-based descriptors and techniques for extracting them from multimedia objects =-=[7, 9, 18, 21]-=-. This approach allows similarity queries based on perceptive considerations [1]. The drawback of that approach lies in the fact that the similarity between images is exclusively evaluated in terms of...

Citation Context

...ng the distances on multiple visual features, such as, color and texture [6], named feature aggregation. We propose a new way to perform feature aggregation, instead of existing heuristic methods [7]–=-=[11]-=-. Let a query image as the prototype, a new feature dissimilarity space is constructed in which an image is represented using the feature distances to the prototype. Then, feature aggregation can be f...

Citation Context

...as MARS [8] and BlobWorld [2] attempting to address this problem using the Boolean logic. To overcome the limit of traditional Boolean logic, decision fusion scheme using fuzzy logic is introduced in =-=[5]-=-. These efforts have achieved certain success in their applications. The results reported individually are obtained in different conditions and the implication of formulae in various schemes and their...

Citation Context

...rm the feature fusion by stacking the descriptor values as a single, large vector and the images is ranked by calculating the distances between these large vectors in a high dimensional feature space =-=[12]-=-. The early fusion approach usually suffers from the dimensionality arising [11]. The other is called as late fusion, which obtains image similarity through combining multiple feature similarities.Com...

Citation Context

...n enhance the retrieval performance to some extent as the different retrieval performances of different features are considered. However, firstly, a certain amount of training data in needed in [11] and [12], secondly, the learned fusion models are not optimal for each image query as a special feature is not equally importance for different image queries. In [13] and [14], the combined similarity between images is measured using one of the features selected by a feature fusion model expressed with logic operation based on Boolean model. To overcome the limitation of traditional Boolean model, [15] introduced a hierarchical decision fusion framework formulated based on fuzzy logic to extend AND and OR operations in Boolean logic. In [13][14][15], the feature fusion models for different image queries are presented with logic-based expressions and they usually require the users to tune appropriate parameters for the fusion models which requiring the user having a good understanding of the low level feature of the query images. In literate [9], the author proposed a query dependent feature fusion method for image retrieval (which is called as local aggregation function in [9]) based on sup...