Table 3.

Comparison of SOTA NR projection-based PCQA metrics.

Ref.	Distortion	Features	Quality Prediction Model
[16]	Compression	Geometry QP, Texture QP, Texture bitrate per pixel (TBPP)	Mathematical Model
[24]	Compression, donwsampling, geometry noise, color noise	Video spatial and temporal information	CNN and DNN
[25]	Compression	Geometry QP, Texture QP	Dual-branch Transformer
[26]	Compression, downsampling, geometry noise, color noise	H-SCNN generated features	DNN
[27]	Compression, downsampling, geometry noise, color noise	CNN generated features	CNN and DNN
[28]	Compression, bandwidth, resolution, buffer length, allocation algorithm	Blockiness, noise, noise ratio, spatial information, and bandwidth, blur, blur ratio	KNN, LR and Sigmoidal Fitting
[29]	Compression, frame rate, viewing distance	Geometry and attribute QPs, occupancy precision, frame rate, and viewing distance	Gradient Boost Regression
[30]	Compression, quality switch, viewing distance	ITU-T P.1203 model features	Fine-tuned ITU-T P.1203 model