Learned Image Downscaling for Upscaling Using Content Adaptive Resampler

Deep convolutional neural network based image super-resolution (SR) models have shown superior performance in recovering the underlying high resolution (HR) images from low resolution (LR) images obtained from the predefined downscaling methods. In this paper, we propose a learned image downscaling method based on content adaptive resampler (CAR) with consideration on the upscaling process. The proposed resampler network generates content adaptive image resampling kernels that are applied to the original HR input to generate pixels on the downscaled image. Moreover, a differentiable upscaling (SR) module is employed to upscale the LR result into its underlying HR counterpart. By back-propagating the reconstruction error down to the original HR input across the entire framework to adjust model parameters, the proposed framework achieves a new state-of-the-art SR performance through upscaling guided image resamplers which adaptively preserve detailed information that is essential to the upscaling. Experimental results indicate that the quality of the generated LR image is comparable to that of the traditional interpolation based method and the significant SR performance gain is achieved by deep SR models trained jointly with the CAR model. The code is publicly available on: https://github.com/sunwj/CAR.

I. Introduction

As the smartphone cameras are starting to rival or beat DSLR cameras, a large number of ultra high resolution images are produced everyday. However, it is always reduced from its original resolution to smaller sizes that are fit to the screen of different mobile devices and web applications. Thus, it is desirable to develop an efficient image downscaling and upscaling method to make such application more practical and resources saving by only generating, storing and transmitting a single downscaled version for preview and upscaling it to high resolution when details are going to be viewed. Besides, the pre-downscaling and post-upscaling operation also helps to save storage and bandwidth for image or video compression and communication [1]–[4].