English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Due to the increasing number of digital images generated every day, also increases the need to recover information contained in large databases of images quickly and accurately. The applied techniques in image retrieval, in general, make use of features extracted from images and in this feature space make comparisons through similarity metrics. As the focus of these techniques of content retrieval is the use of low level features, there is a problem because the semantic content is unrelated with these low level features. Besides, the relations between objects are not considered. This work addresses this problem also known as semantic gap and intends to apply spatial relations with the goal of reducing the semantic gap impact in the scene analysis. In this work, the spatial relations are expressed in terms of spatial prepositions. Thus, in order to define a preposition from an image it is necessary to segment the objects within the image and by each pair of segmented object establish the topology between them. To perform segmentation, neural networks are used. Considering one image, the neural network provides a probabilistic value for every pixel in the image, therefore, the regions covered by similar values belong to the same class and are considered as being the segmented object. Furthermore, these objects are combined in pairs in order to establish binary relations between them. Relying on the objects and their respective regions, mereotopological relations are established according to Region Connection Calculus (RCC). The definition of these relations is performed through a procedure that intends to assign a pair of objects identified in the image with one of the eight relations defined in RCC. These relations established, based on these procedure, are converted to linguistic prepositions (in English language), through another neural network. Finally, the object information and their respective relations is applied to extend an upper ontology. The applied ontology is the Suggested Upper Merged Ontology (SUMO). The procedure of introducing a formal representation allows the establishment of new relations, unknown a priori. This fact enables image retrieval from relations derived through the existing axioms in the applied formal representation.

Representação e recuperação de imagens por meio de relações espaciais entre objetos