Yipeng Liu

Introduction

The established large-scale PCQA dataset named LS-PCQA includes 104 reference point clouds and more than 22,000 distorted samples. In the dataset, each reference point cloud is augmented with 31 types of impairments at 7 distortion levels. Besides, each distorted point cloud is assigned with a pseudo quality score as its substitute of Mean Opinion Score (MOS).

Reference Point Clouds

104 different point clouds include 28 human body models, 48 animal body models and 28 inanimate objects. The following image illustrates the snapshots of all the reference point clouds for LS-PCQA.

Distortions

In total, 31 distortion types are considered.

Distortion Types	Distortion Types
Additive Gaussian noise	Poisson noise
Color noise	Gaussian geometry shifting
High frequency noise	Uniform geometry shifting
Quantization noise	Local missing
Mean shift (intensity shift)	Local offset
Contrast change	Local rotation
Change of color saturation	Luminance noise
Spatially correlated noise	Poisson Reconstruction
Multiplicative Gaussian noise	GPCC-lossless G and lossy A
Color quantization with dither	GPCC-lossless G and nearlossless A
Octree Compression	GPCC-lossy G and lossy A
Down sample	VPCC-lossy G and lossy A
Saltpepper noise	AVS-limitlossy G and lossy A
Rayleigh noise	AVS-lossless G and limitlossy A
Gamma noise	AVS-lossless G and lossy A
Uniform noise (white noise)

Download

Link for reference point clouds: BaiduNetDisk OneDrive

For a quick test, we supply 930 distorted samples with accurate MOS.

Link: BaiduNetDisk OneDrive

Link for whole dataset with generated pseudo MOS: BaiduNetDisk OneDrive