Focused plane segmentation in integral imaging scene reconstruction

Abstract

Computational methods of Integral Imaging allow for the reconstruction of 3D scenes from information extracted from 2D arrays of elemental images. By combining the information from these images, the scene can be decomposed into different planes, enabling focus on the desired depth. The proposed method segments focused regions at each plane of the reconstructed scene, thereby identifying objects at specific depths. This approach allows for the extraction of information from each plane of interest and significantly reduces the effects of occlusions between planes. As a result, our method enhances the accuracy and efficiency of object segmentation by accurately distinguishing between focused and unfocused regions across different depth planes