Lightweight Image Super-Resolution with ConvNeXt Residual Network

Abstract

Single image super-resolution (SISR) based on convolutional neural networks has been very successful in recent years. However, as the computational cost is too high, making it di cult to apply to resource-constrained devices, a big challenge for existing approaches is to nd a balance between the complexity of the CNN model and the quality of the resulting SR. To solve this problem, various lightweight SR networks have been proposed. In this paper, we propose lightweight and e cient residual networks (IRN), which differ from previous lightweight SR networks that aggregate more powerful features by improving feature utilization through complex layer-connection strategies. The main idea is to simplify feature aggregation by using simple and e cient residual modules for feature learning, thus achieving a good trade-off between the computational cost of the model and the quality of the resulting SR. In addition, we revisit the impact of the activation function in the model and observe that different activation functions have an impact on the performance of the model. The experiment results show that IRN outperforms previous state-of-the-art methods in benchmark tests while maintaining a relatively low computational cost. The code will be available at https://github.com/kptx666/IRN.

A Multi-Attention Feature Distillation Neural Network for Lightweight Single Image Super-Resolution

“…To demonstrate the capacity and superiority of MAFDN, representative lightweight SISR methods and recent proposed state-of-the-art models are compared, i.e., SRCNN [19], PAN [74], RFDN [58], FDIWN [38], DDistill-SR [39], PILN [3], JSNet [70], DiVANet [59], EMASRN [60], DRSAN [40], ESRT [67], MMSR [68], HPUN [66], FDSCSR [64], LESR [62], IRN [63], AFAN [61], and FRN [65]. Overall, these competitors cover attention-based, feature distillationbased, and NAS-based SISR models.…”

“…To make a good balance between the computational cost and reconstruction quality, Zhang et al [63] simplifed feature aggregation by using residual modules for feature learning. Wang et al [64] proposed to reduce repetitive feature information via feature deredundancy and selfcalibration, thus enhancing model efciency.…”