A comparative study between recursive and non-recursive algoritms content search in a dual multimedia databases with images

This is a comparative study for recursive and non-recursive content search algorithms in a dual multimedia databases with medical (endoscopic) and different (natural) images. The recursive algorithms in two stages use the same method of determining similarity at each stage and models to represent color content of images in quantified uniform spaces. The non-recursive algorithms use a single iteration and models to represent color content of images in unevenly quantified color spaces. The performance of the search has been measured according to four parameters: reappeal, precision, the quality of retrieval and the cost of retrieval. These are based on two methods of deciding the similarities between the models of images: Minkowski distance and Jaccard generalized measure. The model of representing the images in multimedia database used here are the normalized color histogram. The color space of the images are RGB reduced to 64 and 125 colors and HSV reduced to 60 and 162 colors. The study was realized in a dual database containing 360 endoscopic images grouped in 23 categories and 280 different images grouped in 10 categories. The results are presented both in tables and graphs.