Deep learning-based video quality assessment can be optimized through knowledge distillation and model compression to create a lightweight yet high-performing metric.
