Digital Human Quality Assessment

Digital human quality assessment (DHQA) focuses on objectively measuring the perceptual quality of digitally rendered humans, crucial for applications like virtual and augmented reality. Current research emphasizes developing both full-reference and no-reference assessment methods, often employing deep learning architectures like Vision Transformers and leveraging features extracted from 2D projections and 3D mesh geometry, sometimes incorporating text-based prompts for zero-shot evaluation. The creation of large-scale, subjective quality assessment databases is a key trend, enabling the benchmarking and improvement of DHQA algorithms and ultimately improving the realism and user experience in applications involving digital humans.