SketchFaceNeRF: Sketch-based Facial Generation and Editing in Neural Radiance Field


Lin Gao1,2 *          Feng-Lin Liu1,2           Shu-Yu Chen1           Kaiwen Jiang1,5           Chunpeng Li1          Yu-Kun Lai3          Hongbo Fu4      



1 Institute of Computing Technology, Chinese Academy of Sciences


2 University of Chinese Academy of Sciences


3 Cardiff University      


4 City University of Hong Kong      


5 Beijing Jiaotong University      


* Corresponding author  




Accepted by Siggraph 2023





Figure: Our SketchFaceNeRF system supports both generation and editing of high-quality facial NeRFs from 2D sketches. As shown in the left half part, given a hand-drawn sketch (top-left corner), photo-realistic rendering results with different appearances are synthesized from scratch. The detailed geometry model and free-view rendering results are shown at the bottom. On the right half part, we show sketch-based editing of facial NeRFs and the corresponding geometry, where original faces and geometry are shown in purple boxes, and the results of two consecutive editing steps are shown in green and orange boxes, respectively. During editing, local regions are modified according to the edited sketches highlighted in red, while the geometry and appearance features in unedited regions are well preserved.






Realistic 3D facial generation based on Neural Radiance Fields (NeRFs) from 2D sketches benefits various applications. Despite the high realism of free-view rendering results of NeRFs, it is tedious and difficult for artists to achieve detailed 3D control and manipulation. Meanwhile, due to its conciseness and expressiveness, sketching has been widely used for 2D facial image generation and editing. Applying sketching to NeRFs is challenging due to the inherent uncertainty for 3D generation with 2D constraints, a significant gap in content richness when generating faces from sparse sketches, and potential inconsistencies for sequential multi-view editing given only 2D sketch inputs. To address these challenges, we present SketchFaceNeRF, a novel sketch-based 3D facial NeRF generation and editing method, to produce free-view photo-realistic images. To solve the challenge of sketch sparsity, we introduce a Sketch Tri-plane Prediction net to first inject the appearance into sketches, thus generating features given reference images to allow color and texture control. Such features are then lifted into compact 3D tri-planes to supplement the absent 3D information, which is important for improving robustness and faithfulness. However, during editing, consistency for unseen or unedited 3D regions is difficult to maintain due to limited spatial hints in sketches. We thus adopt a Mask Fusion module to transform free-view 2D masks (inferred from sketch editing operations) into the tri-plane space as 3D masks, which guide the fusion of the original and sketch-based generated faces to synthesize edited faces. We further design an optimization approach with a novel space loss to improve identity retention and editing faithfulness. Our pipeline enables users to flexibly manipulate faces from different viewpoints in 3D space, easily designing desirable facial models. Extensive experiments validate that our approach is superior to the state-of-the-art 2D sketch-based image generation and editing approaches in realism and faithfulness.


















Online System    PyTorch    Jittor






@article {SketchFaceNeRF2023,
    author = {Gao, Lin and Liu, Feng-Lin and Chen, Shu-Yu and Jiang, Kaiwen and Li, Chun-Peng and Lai, Yu-Kun and Fu, Hongbo},
    title = {SketchFaceNeRF: Sketch-Based Facial Generation and Editing in Neural Radiance Fields},
    journal = {ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH 2023)},
    year = {2023},
    volume = 42,
    pages = {159:1--159:17},
    number = 4