Multifeature extracting CNN with concatenation for image denoising

作者：

Highlights：

• We propose a CNN network consisting of skip connection and multifeature extraction.

• We use different-size convolution kernels to extract multiple features.

• Skip connections are used to convey details and prevent gradients from disappearing.

• Our network can realize blind denoising by using noisy images as training labels.

摘要

•We propose a CNN network consisting of skip connection and multifeature extraction.•We use different-size convolution kernels to extract multiple features.•Skip connections are used to convey details and prevent gradients from disappearing.•Our network can realize blind denoising by using noisy images as training labels.

论文关键词：

论文评审过程：Received 22 May 2019, Revised 5 November 2019, Accepted 5 November 2019, Available online 7 November 2019, Version of Record 14 November 2019.

论文官网地址：https://doi.org/10.1016/j.image.2019.115690