Computing and Informatics

Perceptual quality assessment is a tough task especially in the absence of reference information. No-reference image quality assessment is more challenging than full-reference or reduced reference methods, as the system has to model the different image distortions in the form of a quality score. Most of the approaches are based on handcrafted features which are based on natural scene statistics and are specific to some distortion types. These approaches provide high correlation with human opinion score for datasets containing specific distortions, but they fail to generalize well in scenarios were multiple distortions or real-time distortions are present in images. Deep learning algorithms, on the other hand, demonstrated their abilities to learn expert features with better discriminatory power for various classification and regression tasks. It is a big challenge to use those deep learning methods for image quality assessment as the image datasets with human opinion score are very small and cannot be used effectively to train a deep learning algorithm. We experimented with activations of different deep layers of thirteen pre-trained models and checked for their suitability for the task of no-reference quality assessment. Fine-tuning of these models on quality assessment datasets provided even better performance. A Gaussian process regression model is trained on these activations to perform the quality assessment and it provided state-of-the-art performance. Cross-dataset validation demonstrated its performance further and also provided further prospects of research in this direction.