Home > Published Issues > 2015 > Volume 3, No. 2, December 2015 >

A Scaling Method of Sensitive Objects Based on Loss Constraint Triangle Mesh Deformation

Yongmei Zhang and Li Ma
School of Computer Science, North China University of Technology, Beijing, China

Abstract—At present, most of image scaling methods have problems such as loss of smaller sensitive information, easy deformation of sensitive objects, the paper proposes a scaling method of sensitive objects based on loss constraint triangle mesh deformation. Sensitive objects mainly include buildings, cars and vegetation. First to identify the sensitive objects, calculate the sensitivity value of each object using the area size and position relations of objects in the image, then get the initial triangular mesh of the image using the Delaunay mesh subdivision, and the method combines scaling loss function with triangle mesh deformation. The experiment results show the method can scale image-objects at different proportions according to the size of the sensitivity value, reduce the deformation of high sensitivity area and get better scaling effect.

Index Terms—sensitive objects, Delaunay triangulation subdivision, scaling loss function, image scaling

Cite: Yongmei Zhang and Li Ma, "A Scaling Method of Sensitive Objects Based on Loss Constraint Triangle Mesh Deformation," Journal of Image and Graphics, Vol. 3, No. 2, pp. 77-83, December 2015. doi: 10.18178/joig.3.2.77-83