Stereoscopic video quality assessment and enhancement for improving 3D experience

The number of 3D movies is almost doubling each year. Besides, 3D televisions and cameras have become popular. Like 2D content, the 3D images and videos experience several post- processing stages such as sampling and quantization that each level decreases the Quality of Experience (QoE) of the end-user. Consequently, one of the emerging issues is going to be quality assessment and then enhancing the visual quality of stereo images and videos. The depth as the new information channel makes a stereo image and video quality assessment and enhancement far more complicated than its 2D counterpart. This thesis deals with quality assessment and enhancement of stereoscopic videos. For quality enhancement, we propose new methods for stereo video deblurring, super-resolution, and recomposition using stereo style transfer. Specifically, we propose new CNN and Transformer-based model for stereo video quality assessment and enhancement. In addition, we propose unsupervised models for stereo video retargeting and style transfer.