A multi-order derivative feature-based quality assessment model for light field image

• Main Authors: Tian, Yu, Zeng, Huanqiang, Xing, Lu, Chen, Jing, Zhu, Jianqing, Ma, Kai-Kuang
• Other Authors: School of Electrical and Electronic Engineering
• Format: Article
• Language: English
• Published: 2020
• Subjects: Engineering::Electrical and electronic engineering; Light Field Image; Image Quality Assessment
• Online Access: https://hdl.handle.net/10356/142112
• Institution: Nanyang Technological University

This paper presents an image quality assessment (IQA) model exploring the multi-order derivative feature, called Multi-order Derivative Feature-based Model (MDFM), for evaluating the perceptual quality of light field image (LFI). In our approach, for the input reference and distorted LFIs, the multi-order derivative features are extracted by using the discrete derivative filter to represent the image details in different degrees. Then, the similarities of the extracted derivative features are measured independently. Finally, the weight map is established through the maximum value of the second-order derivative feature of reference and distorted LFIs, which is further utilized to pool the similarity map for generating the final score. Extensive simulation results have demonstrated that the proposed MDFM is more consistent with the perception of the HVS on the evaluation of LFI than the classical and state-of-the-art IQA methods.