Image Quality Assessment via Adaptive Pooling

The pooling operation is the key to image quality assessment (IQA), but the traditional pooling treats local quality map equally. In this paper, we propose a novel method named adaptive pooling for image quality assessment, which explicitly considers importance of different local regions. The adaptive pooling operation assigns different weights to local quality map according to the variance of local regions. The proposed method is verified on two challenging IQA databases (CSIQ and TID 2008 databases), and the results demonstrate that it achieves better results than the state-of-the-art methods.