Light field reconstruction from projection modeling of focal stack

Area of Science:

• Computer Vision
• Optics
• Image Reconstruction

Background:

• Focal stacks capture scene information across different focal planes.
• Reconstructing the object-side light field is crucial for advanced 3D scene understanding.
• Existing methods may have limitations in precision and scope.

Purpose of the Study:

• To develop a method for reconstructing the object-side light field from a focal stack.
• To adapt computerized tomography techniques for light field reconstruction.
• To enable accurate depth and surface estimation from the reconstructed light field.

Main Methods:

• Modeling the focal stack as projections of the light field.
• Deriving filtered-backprojection (FBP) and Landweber iterative schemes.
• Applying computerized tomography (CT) principles for inverse problem solving.

Main Results:

• High-precision light field reconstruction achieved using FBP and Simultaneous-Algebraic-Reconstruction-Technique (SART).
• Successful reconstruction of scene depth and surface from the derived light field.
• Validation of the proposed projection modeling approach.

Conclusions:

• The proposed method effectively reconstructs object-side light fields from focal stacks.
• CT-based algorithms like FBP and SART are suitable for light field inverse problems.
• Accurate light field reconstruction facilitates detailed 3D scene analysis.