Journal
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
Volume 27, Issue 1, Pages 178-189Publisher
IEEE COMPUTER SOC
DOI: 10.1109/TVCG.2019.2930512
Keywords
Feature extraction; Semantics; Task analysis; Generative adversarial networks; Visualization; Lighting; Image segmentation; Convolutional neural network; perceptual awareness; sketch simplification
Categories
Funding
- NSFC [61772206, U1611461, 61472145]
- Guangdong R&D key project of China [2018B010107003]
- Guangdong High-level personnel program [2016TQ03X319]
- Guangdong NSF [2017A030311027]
- Guangzhou key project in industrial technology [201802010027]
Ask authors/readers for more resources
In this paper, a method using multi-layer perceptual loss in deep learning is proposed for simplifying sketches while preserving important global structures and details. The approach involves designing a multi-layer discriminator and learning weights automatically to differentiate sketches and clean lines effectively.
Deep learning has been recently demonstrated as an effective tool for raster-based sketch simplification. Nevertheless, it remains challenging to simplify extremely rough sketches. We found that a simplification network trained with a simple loss, such as pixel loss or discriminator loss, may fail to retain the semantically meaningful details when simplifying a very sketchy and complicated drawing. In this paper, we show that, with a well-designed multi-layer perceptual loss, we are able to obtain aesthetic and neat simplification results preserving semantically important global structures as well as fine details without blurriness and excessive emphasis on local structures. To do so, we design a multi-layer discriminator by fusing all VGG feature layers to differentiate sketches and clean lines. The weights used in layer fusing are automatically learned via an intelligent adjustment mechanism. Furthermore, to evaluate our method, we compare our method to state-of-the-art methods through multiple experiments, including visual comparison and intensive user study.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available