Wonder3D's Evaluation Protocol: Datasets and Metrics

Abstract and 1 Introduction

2. Related Works

2.1. 2D Diffusion Models for 3D Generation

2.2. 3D Generative Models and 2.3. Multi-view Diffusion Models

3. Problem Formulation

3.1. Diffusion Models

3.2. The Distribution of 3D Assets

4. Method and 4.1. Consistent Multi-view Generation

4.2. Cross-Domain Diffusion

4.3. Textured Mesh Extraction

5. Experiments

5.1. Implementation Details

5.2. Baselines

5.3. Evaluation Protocol

5.4. Single View Reconstruction

5.5. Novel View Synthesis and 5.6. Discussions

6. Conclusions and Future Works, Acknowledgements and References

Evaluation Datasets. Following prior research [31, 33], we adopt the Google Scanned Object dataset [13] for our evaluation, which includes a wide variety of common everyday objects. Our evaluation dataset matches that of SyncDreamer [33], comprising 30 objects that span from everyday items to animals. For each object in the evaluation set, we render an image with a size of 256×256, which serves as the input. Additionally, to assess the generalization ability of our model, we include some images with diverse styles collected from the internet in our evaluation.

\ Metrics. To evaluate the quality of the single-view reconstructions, we adopt two commonly used metrics Chamfer Distances (CD) and Volume IoU between ground-truth shapes and reconstructed shapes. Since different methods adopt various canonical systems, we first align the generated shapes to the ground-truth shapes before calculating the two metrics. Moreover, we adopt the metrics PSNR, SSIM [62] and LPIPS [74] for evaluating the generated color images.

\

:::info This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license.

:::

:::info Authors:

(1) Xiaoxiao Long, The University of Hong Kong, VAST, MPI Informatik and Equal Contributions;

(2) Yuan-Chen Guo, Tsinghua University, VAST and Equal Contributions;

(3) Cheng Lin, The University of Hong Kong with Corresponding authors;

(4) Yuan Liu, The University of Hong Kong;

(5) Zhiyang Dou, The University of Hong Kong;

(6) Lingjie Liu, University of Pennsylvania;

(7) Yuexin Ma, Shanghai Tech University;

(8) Song-Hai Zhang, The University of Hong Kong;

(9) Marc Habermann, MPI Informatik;

(10) Christian Theobalt, MPI Informatik;

(11) Wenping Wang, Texas A&M University with Corresponding authors.

:::

\