Face Shape Classification Based on Region Similarity, Correlation and Fractal Dimensions

In this paper, we have presented a new idea for face shape classification based on the three techniques such as facial region similarity, correlation and fractal dimensions. The human face is described by different semantic attributes like shape, color and texture. The semantic description of the object attributes helps in searching the object by its contents. The face shape similarity is described with the semantic attribute of the shape like a circle, ellipse, triangular, rectangular or oval, etc. The searching face image using semantic description is potentially important for the law enforcement agencies. The face shape is approximated by using degree of the fitness of the face shape with the geometrical shapes. The result shows that, the proposed approach based on the facial region matching gives effective results for face shape classification.