Abstract. Humans can associate vision and language modalities and thus generate mental imagery, i.e. visual images, from linguistic input in an environment of unlimited inflowing i...
In this a novel supervised learning method is proposed to map low-level visualfeatures to high-level semantic conceptsfor region-based image retrieval. The contributions of thispa...
Wei Jiang, Kap Luk Chan, Mingjing Li, HongJiang Zh...
Abstract. Image classification could be treated as an effective solution to enable keyword-based semantic image retrieval. In this paper, we propose a novel image classification...
We propose a neural network based method for organizing images for content-based image retrieval. We use spectral histogram features, the histograms of filtered images to capture...
—Effective image retrieval by content from database requires that visual image properties are used instead of textual labels to properly index and recover pictorial data. Retriev...