Real-time per-pixel focusing method for light field rendering

image-based rendering, light field, plenoptic function, computational photography

Light field rendering belongs to image-based rendering methods that do not use 3D models but only images of the scene as the input to render new views. Light field approximation, represented as a set of images, suffers from so-called refocusing artifacts due to different depth values of the pixels in the scene. Without the information about the depth in the scene, a proper focusing of the light field scene is limited to a single focusing distance. The correct focusing method is addressed in this work and a real-time solution for focusing of light field scenes, based on statistical analysis of the pixel values contributing to final image, is proposed. Compared to existing techniques, this method does not need a precomputed or acquired depth information. Memory requirements and streaming bandwidth are reduced and real-time rendering is possible even when using a high resolution light field data, yielding visually satisfactory results. Experimental evaluation of the proposed method implemented on GPU is presented in this paper.