A blind stereoscopic image quality assessor using segmented stacked autoencoders that considers the entire visual perception path

PROJECT TITLE :

A Blind Stereoscopic Image Quality Evaluator With Segmented Stacked Autoencoders Considering the Whole Visual Perception Route

ABSTRACT:

Blind stereoscopic image quality assessment (SIQA) methods currently in use are unable to accurately assess the quality of images. One reason is because they lack deep architectures, and the other is that they are created on a biological basis that is weaker than that of the human visual system. DECOSINE (Deep Edge and Color Signal Integrity Evaluator) is based on the entire visual perception route from the eyes to the frontal cortex, with particular emphasis on the processing of edge and colour signals in retinal ganglion cells and lateral geniculate nucleus. S-SAE is used to mimic the visual cortex's complex and deep structure, which has never been employed for SIQA before. In contrast to SIQA measures that require a significant training period, the S-SAE can be used to supplement the weaknesses of Deep Learning-based metrics. DECOSINE's superiority in terms of prediction accuracy and monotonicity is demonstrated through experiments on prominent SIQA databases. For SIQA, we have demonstrated through experiments that our model of the entire visual perception route and use of S-SAE are effective.

Did you like this research project?

To get this research project Guidelines, Training and Code... Click Here