New publication: "ShapeShifter: 3D Variations Using Multiscale and Sparse Point-Voxel Diffusion" by Pierre Alliez and Nissim Maruani (CVPR 2025)
- Research
Published on March 26, 2025
–
Updated on March 26, 2025
Dates
on the March 25, 2025
Pierre Alliez, 3IA Chairholder, and Nissim Maruani, 3IA PhD, have just published a new paper at CVPR 2025.
Pierre Alliez and Nissim Maruani have just made a new publication entitled "ShapeShifter: 3D Variations Using Multiscale and Sparse Point-Voxel Diffusion".
Abstract: ShapeShifter proposes a new 3D generative model that learns to synthesize shape variations based on a single example. While generative methods for 3D objects have recently attracted much attention, current techniques often lack geometric details and/or require long training times and large resources. Our approach remedies these issues by combining sparse voxel grids and multiscale point, normal, and color sampling within an encoder-free neural architecture that can be trained efficiently and in parallel. We show that our resulting variations better capture the fine details of their original input and can capture more general types of surfaces than previous SDF-based methods. Moreover, we offer interactive generation of 3D shape variants, allowing more human control in the design loop if needed.
This paper is part of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025 (CVPR 2025) which will take place from June 11 to 15 at the Music City Center in Nashville (USA).
Read the whole paper
Abstract: ShapeShifter proposes a new 3D generative model that learns to synthesize shape variations based on a single example. While generative methods for 3D objects have recently attracted much attention, current techniques often lack geometric details and/or require long training times and large resources. Our approach remedies these issues by combining sparse voxel grids and multiscale point, normal, and color sampling within an encoder-free neural architecture that can be trained efficiently and in parallel. We show that our resulting variations better capture the fine details of their original input and can capture more general types of surfaces than previous SDF-based methods. Moreover, we offer interactive generation of 3D shape variants, allowing more human control in the design loop if needed.
This paper is part of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025 (CVPR 2025) which will take place from June 11 to 15 at the Music City Center in Nashville (USA).
Read the whole paper
