Blind tone-mapped video quality assessment using spatiotemporal features

With the proliferation of cutting-edge developments in video acquisition, computational imaging, and display technologies, there has been a marked increase in scholarly and industrial interest in high dynamic range (HDR) videos. Tone mapping operators (TMOs), which transform high dynamic range (HDR) content into low dynamic range (LDR) formats, deliver a practical solution for rendering HDR videos on standard LDR displays, where the user experience strongly depends on how effectively the selected TMO performs. There is an urgent need to devise a tone-mapped video quality assessment method for the purpose of comparing various Tone Mapping Operators. Although subjective experiments can provide a trustworthy methodology, they are not only time-intensive and expensive but also pose challenges when it comes to integrating them into optimization processes. We propose a blind objective quality assessment model for tone-mapped videos that integrates intra-frame information, flickering characteristics, and statistical naturalness within the spatiotemporal domain. The experimental results obtained from the Tone-mapped Video Database (TMVD2017) demonstrate the effectiveness of the proposed algorithm. The proposed method exhibits excellent performance compared to the existing representative video quality assessment (VQA) method.