This study explores the use of semantic descriptions that are both qualitative and spatially expressive for image retrieval. It addresses the question of how a qualitative representation performs in comparison to a more quantitative one by employing a semantic-based model approach. The methodology is based on learning qualitative class descriptions and retrieving them into one of five semantically meaningful classes, such as apple, banana, kiwi, pear, pomegranate etc., by applying different qualitative spatial representations to local semantic concepts in a corpus of natural scenes images. Modern image processing techniques and algorithms support efficient semantic image feature extraction and retrieval. This work focuses on content based retrieval using qualitative knowledge driven semantic modelling retrieval .The experimental result show by using qualitative and spatially expressive semantic descriptions, we can improve the accuracy and efficiency of image retrieval for 3D objects.

Semantic modelling, knowledge based, QKDS,VGG16,PCA