Content-based Image Retrieval using a Metric Free Nearness Measure

The focus of this paper is on applying the metric-free nearness measure to the problem of content-based image retrieval. Previously, a metric free nearness measure was defined in [13] by finding common patterns among disjoint description-based neighbourhoods. Thus, only specific neighbourhoods were considered for comparison. When comparing images, it is possible to find many neighbourhoods for comparison. As a result, a new form of the metric-free nearness measure is introduced that measures the similarity between two sets of neighbourhoods, each of which covers a separate digital image. The results are presented by way of digital images of microfossils, which provide a nice example of the utility of the metric-free nearness measure.