Material Detail

Using Fuzzy DLs to Enhance Semantic Image Analysis

Using Fuzzy DLs to Enhance Semantic Image Analysis

This video was recorded at Semantic and Digital Media Technologies. Research in image analysis has reached a point where detectors can be learned in a generic fashion for a significant number of conceptual entities. The obtained performance however exhibits versatile behaviour, reflecting implications over the training set selection, similarities in visual manifestations of distinct conceptual entities, and appearance variations of the conceptual entities. A factor partially accountable for these limitations relates to the fact that machine learning techniques realise the transition from visual features to conceptual entities based solely on information regarding perceptual features. Hence, a significant part of knowledge is missed. In this paper, we investigate the use of formal semantics in order to benefit from the logical associations between the conceptual entities, and thereby alleviate part of the challenges involved in extracting semantic descriptions. More specifically, a fuzzy DL based reasoning framework is proposed for the extraction of enhanced image descriptions based on an initial set of graded annotations, generated through generic image analysis techniques. Under the proposed reasoning framework, the initial descriptions are integrated at a semantic level, resolving inconsistencies emanating from conflicting descriptions. Furthermore, the descriptions are enriched by means of entailment, resulting in more complete image descriptions. Experimentation in the domain of outdoor images has shown very promising results, demonstrating the added value in terms of accuracy and completeness.

Quality

  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material

Comments

Log in to participate in the discussions or sign up if you are not already a MERLOT member.