Enhanced Linear Perspective using Adaptive Intrinsic Camera Parameters

Monocular depth cues allow us to perceive depth from two-dimensional (2-D) images, and linear perspective is one of the most important monocular depth cues. In order to insert improved depth perception into the images, we propose a new method for enhancing linear perspective based on adaptive intrinsic camera parameters in this paper. The threedimensional (3-D) warping technique is utilized to rearrange textures according to their depth values. Using directional dilation and in-paint methods, we fill holes generated by the 3-D warping operation. Experimental results show that the proposed algorithm can enhance linear perspective and improve depth perception of 2-D images without considerable distortion.