Histogram matching for color detection: A preliminary study

Several similarity measures such as Euclidean Distance, Histogram Intersection, Cosine or Quadratic Distances are usually used to calculate the images similarity rating. However, each of the distances has their own limitations when used in particular situation or application. This paper reports a preliminary study of current similarity measures used based on histogram matching. Few experiments on image and color histograms are also presented and discussed. Images of Oil Palm Fruits are used to estimate for its ripeness. This paper will point out the use of existing similarity measures to find the most suitable and efficient one for that particular application.