Alexander Limonov

Human face 3D reconstruction with handheld single 2D camera on mobile devices

In this paper, we propose a live system for 3D reconstruction of human face using only single 2D camera without any 3D sensor. Lack of feature points and homogeneous skin color leads to low quality and success rate in conventional 3D reconstruction algorithms when applied to human face. Moreover, it requires difficult user interaction. To solve this problem, we adopt the facial shape and appearance model to our 3D reconstruction pipeline. This is the first approach which creates a complete 3D face model that can be directly used for 3D printing or virtual reality applications. Model can be acquired using modern smartphone in less than 40 seconds.

33.4: Energy Based Hole-Filling Technique For Reducing Visual Artifacts In Depth Image Based Rendering

In this paper, we propose a new hole-filling approach for DIBR process. The algorithm adaptively selects background extrapolation basis in order to minimize visual artifacts which usually arise in the disocclusion process. Experimental results show that the proposed algorithm helps to reduce edge distortion during the hole-filling process.