Source codes of the paper entitled "Blind Quality Assessment of Screen Content Images via Edge Histogram Descriptor and Statistical Moments"
Run demo.m to extract the EHDSM features for the sample screen content image "im.bmp"
To evaluate the model on screen content datasets you need to do the following steps:
1. Extract the EHDSM features for each image within the dataset using the feature_extractor.p with the recommended parameters in the demo.m
2. Save the extracted features as a .mat file
3. Run the evaluate_model.m to get the median performance through 1000 random train and test split of the dataset
If you have any questions, do not hesitate to contact me via: [email protected]