Fixed smooth convolutional layer for avoiding checkerboard artifacts in CNNs
This addresses a specific technical issue in CNNs and GANs for researchers and practitioners, but it is incremental as it builds on known artifact problems.
The paper tackles checkerboard artifacts in CNNs by proposing a fixed convolutional layer with controllable smoothness, which prevents these artifacts and improves classification performance across multiple CNNs, including VGG8 and ResNet-101, and enhances image quality in GANs.
In this paper, we propose a fixed convolutional layer with an order of smoothness not only for avoiding checkerboard artifacts in convolutional neural networks (CNNs) but also for enhancing the performance of CNNs, where the smoothness of its filter kernel can be controlled by a parameter. It is well-known that a number of CNNs generate checkerboard artifacts in both of two process: forward-propagation of upsampling layers and backward-propagation of strided convolutional layers. The proposed layer can perfectly prevent checkerboard artifacts caused by strided convolutional layers or upsampling layers including transposed convolutional layers. In an image-classification experiment with four CNNs: a simple CNN, VGG8, ResNet-18, and ResNet-101, applying the fixed layers to these CNNs is shown to improve the classification performance of all CNNs. In addition, the fixed layer are applied to generative adversarial networks (GANs), for the first time. From image-generation results, a smoother fixed convolutional layer is demonstrated to enable us to improve the quality of images generated with GANs.