2: A screenshot of our sketching interface for realistic hairstyle design. The qualitative and quantitative evaluations show the advantages of the proposed system over the existing or alternative solutions. Based on the trained networks and the two sketch completion strategies, we build an intuitive interface to allow even novice users to design visually pleasing hair images exhibiting various hair structures and appearance via freehand sketches. Two efficient methods for sketch completion are proposed to automatically complete repetitive braided parts and hair strokes, respectively, thus reducing the workload of users. To train these networks, we present a new dataset containing thousands of annotated hair sketch-image pairs and corresponding hair mattes. To make the networks in the two stages aware of long-term dependency of strokes, we apply self-attention modules to them. At the second stage, another network is trained to synthesize the structure and appearance of hair images from the input sketch and the generated matte. At the first stage, we train a network to predict a hair matte from an input hair sketch, with an optional set of non-hair strokes. Based on these observations, we present SketchHairSalon, a two-stage framework for generating realistic hair images directly from freehand sketches depicting desired hair structure and appearance. We observe that colored hair sketches already implicitly define target hair shapes as well as hair appearance and are more flexible to depict hair structures than orientation maps. Those solutions usually encode hair structures via orientation maps, which, however, are not very effective to encode complex structures. This not only costs users extra labor but also fails to capture complicated hair boundaries. Existing solutions often require a user-provided binary mask to specify a target hair shape. Recent deep generative models allow real-time generation of hair images from sketch inputs. Our two-stage framework automatically generates both hair mattes (Left-Bottom in each example) and hair images (Right in each example) directly from such sketches. SketchHairSalon : Deep Sketch-based Hair Image Synthesis Chufeng Xiao 1 Deng Yu 1 Xiaoguang Han 2 Youyi Zheng 3 Hongbo Fu 1 РІв‚¬вЂ” 1 School of Creative Media, City University of Hong Kong 2 SSE, The Chinese University of Hong Kong, Shenzhen 3 State Key Lab of CAD&CG, Zhejiang University * Corresponding author Accepted by SIGGRAPH Asia 2021 Р”Р‡ig 1: Our SketchHairSalon system allows users to easily create photo-realistic hair images with various hairstyles (e.g., straight, wavy, braided) from freehand sketches (Left-Top in each example), containing colored hair strokes and non-hair strokes (in black).

Hair sketch