Zucheul | Research

Multi-array cameras have been recently developed with many features such as refocusing after taking the photo, focusing on multiple objects, and combining the multiple images over stereo camera. In particular, it might be able to provide accurate depth information of the captured scene, which is fundamental information for a wide range of 3-D applications. In this project, our goal is to develop an effective disparity enhancement algorithm combining complementary multiple array disparity maps. We consider a 3x3 camera configuration as a basic setup for multi-array system. We create multi-array images and video, using 3D rendering software tool (AUTODESK 3ds Max) and add camera noise to simulate a real camera system. We exploit horizontal/vertical/narrow-baseline/wide-baseline disparity estimation and show how to fuse them.