Color-Separation Framework for Ceramic-Tiles based on Few Labeled Data by HSV and SVM

The color-separation of ceramic-tiles is still dominated by human eye discrimination in actual production. According to the actual production, this paper proposes a ceramic-tile color-separation framework based on few labeled data training, which can realize the color-separation of ceramic-tiles with complex patterns. Firstly, the image with ceramic-tile only is obtained by image pre-processing, then the features are extracted by histogram statistics under HSV color space in sub-regions, and finally the support vector machine model is trained to obtain the classifier using the features extracted from a small number of training data and this classifier is applied to the automatic color-separation of ceramic-tiles. Experimental results of simulation and Practical application have proven the effectiveness of the proposed method.