Image Super-resolution with An Enhanced Group Convolutional Neural Network

Abstract

CNNs with strong learning abilities are widely chosen to resolve super-resolution problem. However, CNNs depend on deeper network architectures to improve performance of image superresolution, which may increase computational cost in general. In this paper, we present an enhanced super-resolution group CNN (ESRGCNN) with a shallow architecture by fully fusing deep and wide channel features to extract more accurate low-frequency information in terms of correlations of different channels in single image super-resolution (SISR). Also, a signal enhancement operation in the ESRGCNN is useful to inherit more long-distance contextual information for resolving long-term dependency. An adaptive up-sampling operation is gathered into a CNN to obtain an image super-resolution model with low-resolution images of different sizes. Extensive experiments report that our ESRGCNN surpasses the state-of-the-arts in terms of SISR performance, complexity, execution speed, image quality evaluation and visual effect in SISR. Code is found at https://github.com/hellloxiaotian/ESRGCNN.

Lightweight spatial-channel adaptive coordination of multilevel refinement enhancement network for image reconstruction

“…Gao [27] proposed wide residual distillation connection, which connects the features of different convolution stages in the module while maintaining the light weight of the network, so as to realize the information exchange of different scale features. Tian [3] alternately uses ordinary convolution and lightweight grouped convolution to achieve feature extraction and multi-stage feature aggregation. In addition, [28], [29], [30] also achieve excellent reconstruction performance with a small number of parameters and computational complexity.…”

“…In recent years, with the rapid development of deep learning, the reconstruction method based on deep learning has been widely concerned by researchers. With the powerful nonlinear characterization ability of convolution neural networks, researchers have proposed many reconstruction algorithms with excellent performance, such as DNCL [1], FilterNet [2], ESRGCNN [3], etc. In 2016, Dong [4] constructed a shallow neural network containing three convolution layers to learn the mapping relationship between LR and HR, and obtained more excellent reconstruction results than traditional algorithms.…”

Semantic Segmentation of Side-Scan Sonar Images with Few Samples

“…Image segmentation models represented by FCN [9], U-net [10], PSPNet [11] have also attracted the attention of many researchers. GAN networks are also widely used in machine learning data generation to solve the problem of insufficient data [12][13][14]. Given the good results of these algorithms, the researchers hope to apply them to underwater acoustic images, thereby advancing the field of underwater sensing and detection.…”