COMPUTATIONAL MODEL FOR COLOR CLASSIFICATION BASED ON KNEAREST NEIGHBOR COLOR MINING TECHNIQUE

In this thesis, we are developed a effective and accurate methodology for identification and recognition of human vision system for example camera and robots both are performed the colour classifications action in practical life. The objective of the proposed methodology is colour classification of the image based on hybrid concepts of K-NN data mining technique and pixel processing. K-NN technique is one of the vital classification approach of data mining because it is analyzed the pixel with neighbour pixel with 8 connected quantifiable manner of the image for colour mining on the behalf of K=8 value. We know that pixel processing is a essential part of image processing. In colour mining we calculated particular RGB colour with its combinations value like Red Colour and its all possible combinations and after that analysed pixels and calculated some variable like that colour variance, maximum colour, minimum colour, median colour, area of colour and also colour quality by using MATLAB tools.