text stringlengths 476 13.7k |
|---|
Instant 3D Human Avatar Generation using Image Diffusion Models Nikos Kolotouros, Thiemo Alldieck, Enric Corona, Eduard Gabriel Bazavan, and Cristian Sminchisescu Google Research⋆ {kolotouros,alldieck,egbazavan,enriccorona,sminchisescu}@google. com Fig. 1:We present Avatar Pop Up, a new method for the automatic generat... |
2 N. Kolotouros et al. 1 Introduction We present Avatar Pop Up, a method for instant generation of rigged full-body 3D human avatars, with multimodal controls in the form of text, images, and/or human pose and shape. The remarkable recent progress in image synthesis [15,24,54,56,58,63] acted as a catalyst for a wide ra... |
Instant 3D Human Avatar Generation using Image Diffusion Models 3 Control from Text Control from Image Body Pose/Shape Control Generates Geometry Generates Texture Editable Runtime ✓ ✓ ✓ ✓ ✓ ✓ /✗several hours Optimization-based text-to-3D [26,32,37] ✗ ✓ ✗ ✓ ✓ /✗ ✗ seconds Image-to-3D [4,8,60,61] ✓ ✗ ✓ ✓ ✗ ✗ minutes Hyb... |
4 N. Kolotouros et al. are diverse in terms of shape, appearance, clothing and various accessories. In these works, the optimization process is often regularized using a 3D body model [32,37,78], which also enables animation. However, such approaches gen-erally take hours per instance, and rendering is slow. With the a... |
Instant 3D Human Avatar Generation using Image Diffusion Models 5 <latexit sha1_base64="NADb Mm ZPYq63m CRXHRxo QZz Hn DI=">AAAB/Hicb VBLSw Mx GMz WV62v1R69BIt QQcqu+Do Wv Xis YB/QLi Wbzbah2WRJss Ky1L/ix YMi Xv0h3vw3Zts9a Ot Ay DDzf WQyfsyo0o7zb ZVWVtf WN8qbla3tnd09e/+go0Qi MWljw YTs+Ug RRjlpa6o Z6c WSo Mhnp Ot Pbn O/+... |
6 N. Kolotouros et al. 3. 1 Controllable Text-to-Image Generator Recent advances in diffusion-based text-to-image generation networks [56] have enabled synthesizing high-quality images given only a text prompt as input. However, for certain use cases, such as human generation, it is difficult to inject fine-grained, in... |
Instant 3D Human Avatar Generation using Image Diffusion Models 7 No text conditioning + “gray hair ” + “with a tattoo ” Fig. 3: Diverse back view hypotheses. Conditioned on the front view, our method is able to generate diverse plausible back views of the person, with different hairstyles, wrinkle patterns, or lightin... |
8 N. Kolotouros et al. methods. Specifically, given a collection of input image signals I={If, Ib,G}, we first concatenate them, and then use a convolutional encoder Gto compute a pixel-aligned feature map G(I). The 3D body control signal Gis optional and may be omitted, e. g. for single-image reconstruction. Then, eac... |
Instant 3D Human Avatar Generation using Image Diffusion Models 9 R-Prec. ↑Top-3 ↑R-Prec. ↑Top-3 ↑Runtime ↓ 0. 68 0. 92 0. 040. 25 8h Dream Human [32] 0. 56 0. 82 0. 03 0. 15 3h TADA [37]--0. 03 0. 08 3m CHUPA [31] 0. 58 0. 73 0. 08 0. 17 ∼2s Ours 0. 62 0. 770. 11 0. 17 ∼10s Ours (high quality)Color Geometry Table 2: N... |
10 N. Kolotouros et al. Fig. 4: Diversity of our 3D generation. For the same text prompt and the same pose and shape conditioning, our model can generate a diverse set of 3D avatars that respect both the text and the 3D body controls. Ch. ↓Io U↑NC↑Color. F Color B. Normal F. Normal B. 7. 10. 500. 72--29. 6 57. 2 PIFu [... |
Instant 3D Human Avatar Generation using Image Diffusion Models 11 A person wearing a chef's hat A person wearing a black off-shoulder top and a denim skirt Our s T AD A D r e am H um an CHUP A Fig. 5: Comparisons with text-to-3d human generation methods. Our method generates high quality results that respect the text ... |
12 N. Kolotouros et al. Ours PHORHUM PIFUHD Fig. 6: Qualitative comparisons with state-of-the-art single image 3D re-construction methods. Our method produces front color and normals on par with state-of-the-art and much more detailed back view hypotheses. Ours T e CH Fig. 7:Additionalcomparisonswith Te CH. Te CHisopti... |
Instant 3D Human Avatar Generation using Image Diffusion Models 13 Fig. 8: Reposing example.. We first reconstruct an avatar “ wearing a gray suit ” in the A-pose and then we transfer it to different poses. reconstruction network during generation, where we have control over the target pose. To do so, we use the same s... |
14 N. Kolotouros et al. “Wearing a green shirt” “Wearing a suit” “Wearing a purple shirt and gray jacket” “Wearing a black shirt and khaki pants” Fig. 9: Identity preserving 3D avatar editing. Our method allows for editing the clothing, while preserving the identity of the generated person. In each row we're using the ... |
Instant 3D Human Avatar Generation using Image Diffusion Models 15 References 1. https://renderpeople. com/ 2. Abdal, R., Yifan, W., Shi, Z., Xu, Y., Po, R., Kuang, Z., Chen, Q., Yeung, D. Y., Wetzstein, G. : Gaussian shell maps for efficient 3d human generation. ar Xiv preprint ar Xiv:2311. 17857 (2023) 3. Al Bahar, B... |
16 N. Kolotouros et al. 19. Han, X., Cao, Y., Han, K., Zhu, X., Deng, J., Song, Y. Z., Xiang, T., Wong, K. Y. K. : Headsculpt: Crafting 3d head avatars with text. Advances in Neural Information Processing Systems 36(2024) 20. Han, X., Wu, Z., Wu, Z., Yu, R., Davis, L. S. : Viton: An image-based virtual try-on network. ... |
Instant 3D Human Avatar Generation using Image Diffusion Models 17 37. Liao, T., Yi, H., Xiu, Y., Tang, J., Huang, Y., Thies, J., Black, M. J. : Tada! text to animatable digital avatars. In: 3DV (2023) 38. Lin, C. H., Gao, J., Tang, L., Takikawa, T., Zeng, X., Huang, X., Kreis, K., Fidler, S., Liu, M. Y., Lin, T. Y. : ... |
18 N. Kolotouros et al. 54. Ramesh, A., Pavlov, M., Goh, G., Gray, S., Voss, C., Radford, A., Chen, M., Sutskever, I. : Zero-shot text-to-image generation. In: International Conference on Machine Learning. pp. 8821-8831. PMLR (2021) 55. Ramesh, A., Pavlov, M., Goh, G., Gray, S., Voss, C., Radford, A., Chen, M., Sutskev... |
Instant 3D Human Avatar Generation using Image Diffusion Models 19 72. Xu, H., Bazavan, E. G., Zanfir, A., Freeman, W. T., Sukthankar, R., Sminchisescu, C. : Ghum & ghuml: Generative 3D human shape and articulated pose models. In: CVPR (2020) 73. Xu, Y., Yang, Z., Yang, Y. : Seeavatar: Photorealistic text-to-3d avatar ... |
README.md exists but content is empty.
- Downloads last month
- 8