Auto-Encoded Supervision for Perceptual Image Super-Resolution

This work tackles the fidelity objective in the perceptual super-resolution (SR) task. Specifically, we address the shortcomings of pixel-level L p loss (L pix ) in the GAN-based SR framework. Since L pix is known to have a trade-off relationship against perceptual quality, prior methods often multiply a small scale factor or utilize low-pass filters. However, this work shows that these circumventions fail to address the fundamental factor that induces blurring. Accordingly, we focus on two points: 1) precisely discriminating the subcomponent of L pix that contributes to blurring, and 2) only guiding based on the factor that is free from this trade-off relationship. We show that they can be achieved in a surprisingly simple manner, with an Auto-Encoder (AE) pretrained with L pix . Accordingly, we propose the Auto-Encoded Supervision for Optimal Penalization loss (L AESOP ), a novel loss function that measures distance in the AE space 1 , instead of the raw pixel space. By simply substituting L pix with L AESOP , we can provide effective reconstruction guidance without compromising perceptual quality. Designed for simplicity, our method enables easy integration into existing SR frameworks. Extensive experiments demonstrate the effectiveness of AESOP.