HR Human: Modeling Human Avatars with Triangular Mesh and High-Resolution Textures from Videos
International Conference on Computational Visual Media(2025)
Abstract
Recently, implicit neural representation has been widely used to generateanimatable human avatars. However, the materials and geometry of thoserepresentations are coupled in the neural network and hard to edit, whichhinders their application in traditional graphics engines. We present aframework for acquiring human avatars that are attached with high-resolutionphysically-based material textures and triangular mesh from monocular video.Our method introduces a novel information fusion strategy to combine theinformation from the monocular video and synthesize virtual multi-view imagesto tackle the sparsity of the input view. We reconstruct humans as deformableneural implicit surfaces and extract triangle mesh in a well-behaved pose asthe initial mesh of the next stage. In addition, we introduce an approach tocorrect the bias for the boundary and size of the coarse mesh extracted.Finally, we adapt prior knowledge of the latent diffusion model atsuper-resolution in multi-view to distill the decomposed texture. Experimentsshow that our approach outperforms previous representations in terms of highfidelity, and this explicit result supports deployment on common renderers.
MoreTranslated text
PDF
View via Publisher
AI Read Science
Must-Reading Tree
Example

Generate MRT to find the research sequence of this paper
Data Disclaimer
The page data are from open Internet sources, cooperative publishers and automatic analysis results through AI technology. We do not make any commitments and guarantees for the validity, accuracy, correctness, reliability, completeness and timeliness of the page data. If you have any questions, please contact us by email: report@aminer.cn
Chat Paper
Summary is being generated by the instructions you defined