Table 2.

Comparison of SOTA NR model-based PCQA metrics.

Ref.	Distortion	Geometry Features	Color Features	Feature Statistics	Quality Prediction Model
[19]	Compression, downsampling, geometry noise, color noise	3D coordinates, occupation index	R, G, B color channels	-	Sparse CNN
[20]	Compression, downsampling, geometry noise, color noise	Point-level: curvature, anisotropy, linearity, planarity, and sphericity	L, A, B color channels	Mean, standard deviation, entropy, GGD, AGGD, Gamma	SVR
[21]	Compression, downsampling, geometry noise, color noise	PC structure	PC texture	-	Graph convolutional network
[22]	Compression, downsampling, geometry noise, color noise	Geometric distance, mean curvature	Gray-level features	-	CNN