Subjective video quality assessment generally comes across with semantically labeled evaluation scales (e.g. Excellent, Good, Fair, Poor and Bad on a single stimulus, 5 level grading scale). While suspicions about an eventual bias these labels induce in the quality evaluation always occur, to the best of our knowledge, very few state-of-the-art studies target an objective assessment of such an impact. Our study presents a neural network solution in this respect. We designed a 5-class classifier, with 2 hidden layers, and a softmax output layer. An ADAM optimizer coupled to a Sparse Categorical Cross Entropy function is subsequently considered. The experimental results are obtained out of processing a database composed of 440 observers scoring about 7 hours of video content of 4 types (high-quality stereoscopic video content, low-quality stereoscopic video content, high-quality 2D video, and low-quality 2D video). The experimental results are discussed and confrontment to the reference given by a probability-based estimation method. They show an overall good convergence between the two types of methods while pointing out to some inner applicative differences that are discussed and explained.
In recent years, light field technology has attracted the interest of academia and industry, thanks to the possibility of rendering 3D scenes in a more realistic and immersive way. In particular, light field displays have been consistently investigated for their ability to offer a glass-free 3D viewing experience. Among others, tensor displays represent a promising way to render light field contents. However, only a few prototypes of such type of displays have been implemented and are available to the scientific community. As a direct consequence, the visual quality of such displays has not been rigorously investigated. In this paper, we propose a new framework to assess the visual quality of light field tensor displays on conventional 2D screens. The multilayer components of the tensor displays are virtually rendered on a typical 2D monitor through the use of a GUI, and different viewing angles can be accessed by simple mouse interactions. Both single and double stimulus methodologies for subjective quality assessment of light field contents are supported in this framework, while the total time of interaction is recorded for every stimulus. Results obtained in two different laboratory settings demonstrate that the framework can be successfully used to perform subjective quality assessment of different compression solutions for light field tensor displays.