Enhanced dual branches network for arbitrary‐scale image super‐resolution

Abstract

Deep convolutional neural networks (CNNs) are of great improvement for single image super-resolution (SISR). However, most existing SISR pre-trained models can only perform single image restoration and the upscale factors cannot be non-integers, which limits its application in real-world scenarios. In this letter, an enhanced dual branches network (EDBNet) in upsampling network is proposed to generate arbitraryscale super-resolution (SR) images. Specifically, the authors design a scale-guidance upsampling module (SGU) by adding the scale factors and pixel-level features to guide the weights of convolution. The SGU module performs discriminant learning for each instance in the same batch. Extensive experiments on four benchmark datasets show that the proposed method can achieve superior SR results.

Multiregression spatially variant blur kernel estimation based on inter‐kernel consistency

“…Introduction: Single-image super-resolution (SR) models [1][2][3] have demonstrated excellent image restoration performance. In image processing fields, such as display technology, image sensors, video processing, and human visual systems, single-image SR models are extensively used to enhance performance.…”