DeepSDF
Introduction
DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation
- 数据集
- ShapeNet
- 代码
- C++\python
- 学习连续的有符号距离函数（Signed Distance Function，SDF）
- SDF的零极线，差不多是为0的等势面
- 输入：
- 3D点云 + 噪声或缺失
- 用于训练SDF，并构建3D模型
- 模型通过输入的部分和嘈杂的3D数据点，使用神经网络进行训练，并生成一个能够表示形状表面的连续距离函数。这个距离函数可以通过计算一个点到形状边界的距离，并根据符号指示区域是形状内部还是外部。通过学习这样的连续距离函数，模型能够实现对形状的表示、重建、插值和完成任务。
- DeepSDF通过神经网络将未知形状的表面表示为SDF的零级线，并可用于插值和完成部分和噪声3D数据。相比之前的方法，DeepSDF在学习的3D形状表示和完成任务上取得了最先进的性能。这项工作的关键创新之处在于引入了自动解码器（auto-decoder）的学习方案，将形状编码和网络参数联合进行优化，同时在数据准备、网络架构和实验设计等方面进行了深入的探索和分析。
The authors normalized each mesh to a unit sphere and sampled singed distance values for 500,000 spatial points. To capture greater geometric detail of the objects for better model training, sampling was most aggressive near surface areas. Compared with traditional auto-encoders with both encoder and decoder architectures, such decoder-only networks make the model more compact while maintaining good performance.
- 直接用隐变量去Auto-Decoder
Training a specific neural network for each shape is neither feasible nor very useful. Instead, we want a model that can represent a wide variety of shapes, discover their common properties, and embed them in a low dimensional latent space. To this end, we introduce a latent vector z, which can be thought of as encoding the desired shape, as a second input to the neural network as depicted in Fig. 3b.
Notice
git
Shape Completion The current release does not include code for shape completion. Please check back later!
