PHD Logo

PHD
Personalized 3D Human Body Fitting with Point Diffusion

ICCV 2025

(Equal Advisory)
ETH Zürich Logo 1ETH Zürich
AI center Logo 2ETH AI Center
Meta Logo 3Meta
Paper Code Video
Teaser Video

A compelling one-sentence caption describing the amazing result shown above.

Abstract

This is where your abstract goes. It should provide a concise summary of your work. Briefly introduce the problem, your proposed method, key results, and the main contribution. Keep it engaging and accessible to a broad audience. For example, you could start with: "We introduce a novel method for... that addresses the long-standing challenge of... Our approach, which we call 'MethodName', leverages a unique architecture to... We demonstrate state-of-the-art performance on several challenging benchmarks, significantly outperforming previous methods in..."

Method

Here, you can provide a high-level overview of your method. Explain the core intuition behind your approach without getting lost in excessive detail (that's what the paper is for). A diagram is highly recommended to illustrate the pipeline or architecture.

Method Diagram

Fig 1: An overview of our proposed method. Input data is processed through several stages to produce the final output.

Results

Showcase your best results here. Use a gallery of images, GIFs, or videos. Side-by-side comparisons with prior work can be very effective.

Result 1

Caption for the first result, perhaps comparing it to a baseline method.

Result 2

Caption for the second result, highlighting a specific feature or success case.

Result 3

Another impressive result. GIFs work great here to show dynamic processes.

Result 4

This result could demonstrate the robustness of your method on a difficult input.

Citation

If you find our work useful, please consider citing:

@inproceedings{ho2025phd, title={PHD: Personalized 3D Human Body Fitting with Point Diffusion}, author={Ho, Hsuan-I and Guo, Chen and Wu, Po-Chen and Shugurov, Ivan and Tang, Chengcheng and Mittal, Abhay and An, Sizhe and Kaufmann, Manuel and Zhang, Linguang}, booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, year={2025} }