Close-up View synthesis by Interpolating Optical Flow

TL;DR

This paper introduces a novel method for synthesizing close-up virtual views using only optical flow, enabling pseudo 3D effects and magnified views without depth sensors, improving visual fidelity in virtual navigation.

Contribution

It proposes a bidirectional optical flow interpolation technique and optical-flow-value application to generate high-fidelity close-up views without depth information.

Findings

Achieves clear magnified views with reduced distortion.

Enables viewpoint interpolation without depth sensors.

Improves visual quality in virtual navigation systems.

Abstract

The virtual viewpoint is perceived as a new technique in virtual navigation, as yet not supported due to the lack of depth information and obscure camera parameters. In this paper, a method for achieving close-up virtual view is proposed and it only uses optical flow to build parallax effects to realize pseudo 3D projection without using depth sensor. We develop a bidirectional optical flow method to obtain any virtual viewpoint by proportional interpolation of optical flow. Moreover, with the ingenious application of the optical-flow-value, we achieve clear and visual-fidelity magnified results through lens stretching in any corner, which overcomes the visual distortion and image blur through viewpoint magnification and transition in Google Street View system.