Performance and Analysis of the Automated Semantic Object and Spatial Relationships Extraction in Traffic Images

Abstract

Extraction and representation of spatial relations semantics among objects are important as it can convey important information about the image and to further increase the confidence in image understanding which contributes to richer querying and retrieval facilities. This paper discusses the performance of the automated object spatial relationships semantic information extraction as proposed. Experiments have been conducted to demonstrate that the proposed automated object spatial relationship semantic extraction is succeeded to capture the semantic spatial relationship features in the images.