Reasoning as Representation: Rethinking Visual Reinforcement Learning in Image Quality Assessment

TL;DR

Reasoning-based image quality assessment (IQA) models trained through reinforcement learning (RL) exhibit exceptional generalization, yet the underlying mechanisms and critical factors driving this capability remain underexplored in current research. Moreover, despite their superior performance, these models incur inference energy usage and late...

Opening excerpt from the authors’ abstract. source