Anamorphic Image Generation Using Hybrid Texture Syn- thesis

Anamorphic image is an original image by special distortion, the observer can not get any information from the image at first glance. The original image can be see only the deformation of the image projected back to the specific shape of the mirror surface. Due to anamorphic image presented visual effects is so interesting and attractive, there are lots of artists committed to the anamorphic art invention from the early Renaissance. Different from the general image deformation, artists decorated with small pieces of pattern in the original meaningless anamorphic image, to create a visual double anamorphic art. However, the creation process of anamorphic art would be tedious, the skill always hold to professional artists can not be popular. For the reason, in this paper we propose a semi-automated system, users only need to input an original image and a small image database, the system will automatically analyze the characteristics of original image with a small image, and using the partial shape matching algorithm according to the feature analysis, and automatically selects the matching images to fill the anamorphic image to achieve the artist's decorative skill. Finally, we use the luminance optimization algorithm, adjust the small image's gradation, resulting in the final anamorphic image. The results show that the proposed algorithm not only to maintain the original small image can be recognizable in the anamorphic image but also clearly restore the original image through the ray tracing method to achieve the visual duality anamorphic image.