SuperStyleNet: Deep Image Synthesis with Superpixel Based Style Encoder

TL;DR

SuperStyleNet introduces a superpixel-based style encoder for image synthesis, effectively capturing local image details and spatial relationships, leading to improved visual quality and enabling detailed spatial style editing.

Contribution

The paper presents a novel superpixel-based style encoder that preserves local image information and spatial relationships, enhancing image synthesis quality and editing capabilities.

Findings

Outperforms state-of-the-art methods in visual quality

Enables detailed spatial style editing

Achieves high-quality image synthesis

Abstract

Existing methods for image synthesis utilized a style encoder based on stacks of convolutions and pooling layers to generate style codes from input images. However, the encoded vectors do not necessarily contain local information of the corresponding images since small-scale objects are tended to "wash away" through such downscaling procedures. In this paper, we propose deep image synthesis with superpixel based style encoder, named as SuperStyleNet. First, we directly extract the style codes from the original image based on superpixels to consider local objects. Second, we recover spatial relationships in vectorized style codes based on graphical analysis. Thus, the proposed network achieves high-quality image synthesis by mapping the style codes into semantic labels. Experimental results show that the proposed method outperforms state-of-the-art ones in terms of visual quality and quantitative measurements. Furthermore, we achieve elaborate spatial style editing by adjusting style codes.