Image Generation with Self Pixel-wise Normalization
This work addresses a problem for researchers and practitioners in image generation by enabling pixel-wise normalization without paired masks, though it is incremental as it builds on existing GAN-based techniques.
The paper tackled the limitation of region-adaptive normalization methods in image generation, which require mask images, by proposing self pixel-wise normalization (SPN) that uses a self-latent mask to perform pixel-adaptive affine transformations without external data, resulting in significant improvements in Frechet inception distance (FID) and Inception score (IS) across various datasets.
Region-adaptive normalization (RAN) methods have been widely used in the generative adversarial network (GAN)-based image-to-image translation technique. However, since these approaches need a mask image to infer the pixel-wise affine transformation parameters, they cannot be applied to the general image generation models having no paired mask images. To resolve this problem, this paper presents a novel normalization method, called self pixel-wise normalization (SPN), which effectively boosts the generative performance by performing the pixel-adaptive affine transformation without the mask image. In our method, the transforming parameters are derived from a self-latent mask that divides the feature map into the foreground and background regions. The visualization of the self-latent masks shows that SPN effectively captures a single object to be generated as the foreground. Since the proposed method produces the self-latent mask without external data, it is easily applicable in the existing generative models. Extensive experiments on various datasets reveal that the proposed method significantly improves the performance of image generation technique in terms of Frechet inception distance (FID) and Inception score (IS).