Visual Quality Assessment of Point Clouds Compared to Natural Reference Images

This paper proposes a point cloud (PC) visual quality assessment (VQA) framework that reflects the human system (HVS). The proposed compares natural images acquired using digital camera and PC generated via 2D projection in terms of appropriate objective evaluation metrics. Humans primarily consume images; thus, knowledge is typically formed from images. Thus, can be more reliable reference data than data. performs an image alignment process based on feature matching warping to use as which enhances similarities corresponding facilitates identifying VQA metrics used reflect HVS effectively. We constructed database three qualities, subjective VQAs were conducted. experimental result demonstrates acceptable consistency among different qualities appears compare global structural similarity found SSIM, MAD, GMSD achieved remarkable Spearman rank-order correlation coefficient scores 0.882, 0.871, 0.930, respectively. by comparing between