Stitched Wide Field of View Light Field Image Quality Assessment: Benchmark Database and Objective Metric

Due to the limitation of commercial light field camera hardware devices, the imaging field of view is quite narrow. Numerous Light Field Image (LFI) stitching algorithms have been developed to expand the field of view. However, it is highly challenging to compare the performance of LFI stitching algorithms in a fair manner. Currently, due to the absence of a comprehensive benchmark database for subjective rating and a reliable objective quality metric, it is fairly difficult to comprehensively and accurately compare the actual performance of existing LFI stitching algorithms. In this study, we dedicate our efforts to the development of quality metrics for stitched Wide field of view LFI (WLFI) from subjective and objective assessment aspects. Specifically, we build the first stitched WLFI database, which provides the stitched WLFIs generated by eight representative LFI stitching algorithms, along with their corresponding subjective rating scores. Secondly, an effective blind stitched WLFI quality metric is developed to accurately assess the visual quality degradation. Extensive experiments conducted over our established WLFI database demonstrate that the proposed metric achieves higher consistency with subjective ratings than the competing quality metrics.

I. Introduction

With the rapid development of emerging immersive media, the applications of Light Field Image (LFI) are becoming increasingly widespread, such as view rendering, post-capture refocusing, occlusion rendering, and free viewpoint video services [1], [2], [3], [4]. However, LFIs still face numerous challenges in practical applications. For example, the volume of data in an LFI is quite huge, which brings significant difficulties for storage and transmission. Typically, performing sparse viewpoint sampling and compression operations on LFI is essential, but LFI compression and reconstruction processing inevitably lead to its quality degradation. To address these issues, several high-performance quality metrics have been developed [5], [6], [7], [8], [9]. Another issue that needs to be noted is that light field cameras can only capture the scene content with narrow Field Of View (FOV) due to the limitations of hardware devices, for example, the imaging FOV of some light field camera is only with 60°×43°. In other words, during the process of LFI acquisition, light field cameras are unable to capture wide FOV scene content in one shot. Presently, there exist two common schemes to expand the FOV: 1) The hardware system of the light field camera should be modified and upgraded, which is relatively costly, technically challenging, and may potentially result in increased camera size; and 2) From the perspective of computational photography, LFI stitching technologies are utilized to expand the FOV. Relatively speaking, the latter choice is relatively simple, efficient, and easy to implement. Consequently, LFI stitching is more widely used in actual applications.