# DOPE: Distillation Of Part Experts for whole-body 3D pose estimation in the wild

arXiv: https://arxiv.org/abs/2008.09457 · HTML: https://mlsanity.com/paper/2008.09457/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Aug 2020.

## Built on by (most cited first, 25 of 25)

- [Effective Whole-body Pose Estimation with Two-stages Distillation](https://mlsanity.com/paper/2307.15880/) (2023)
- [Revisiting Skeleton-based Action Recognition](https://mlsanity.com/paper/2104.13586/) (2021)
- [Collaborative Regression of Expressive Bodies using Moderation](https://mlsanity.com/paper/2105.05301/) (2021)
- [Recovering 3D Human Mesh from Monocular Images: A Survey](https://mlsanity.com/paper/2203.01923/) (2022)
- [PoseGPT: Quantization-based 3D Human Motion Generation and Forecasting](https://mlsanity.com/paper/2210.10542/) (2022)
- [Human-Aware Object Placement for Visual Environment Reconstruction](https://mlsanity.com/paper/2203.03609/) (2022)
- [Online Knowledge Distillation for Efficient Pose Estimation](https://mlsanity.com/paper/2108.02092/) (2021)
- [Body Meshes as Points](https://mlsanity.com/paper/2105.02467/) (2021)
- [Data-Free Knowledge Transfer: A Survey](https://mlsanity.com/paper/2112.15278/) (2021)
- [H3WB: Human3.6M 3D WholeBody Dataset and Benchmark](https://mlsanity.com/paper/2211.15692/) (2022)
- [Vision-Based Human Pose Estimation via Deep Learning: A Survey](https://mlsanity.com/paper/2308.13872/) (2023)
- [SHOWMe: Benchmarking Object-agnostic Hand-Object 3D Reconstruction](https://mlsanity.com/paper/2309.10748/) (2023)
- [PoseBERT: A Generic Transformer Module for Temporal 3D Human Modeling](https://mlsanity.com/paper/2208.10211/) (2022)
- [Human-M3: A Multi-view Multi-modal Dataset for 3D Human Pose Estimation in Outdoor Scenes](https://mlsanity.com/paper/2308.00628/) (2023)
- [CondiMen: Conditional Multi-Person Mesh Recovery](https://mlsanity.com/paper/2412.13058/) (2024)
- [Voice2Mesh: Cross-Modal 3D Face Model Generation from Voices](https://mlsanity.com/paper/2104.10299/) (2021)
- [Learning Joint Representation of Human Motion and Language](https://mlsanity.com/paper/2210.15187/) (2022)
- [Encapsulating Knowledge in One Prompt](https://mlsanity.com/paper/2407.11902/) (2024)
- [HOSt3R: Keypoint-free Hand-Object 3D Reconstruction from RGB images](https://mlsanity.com/paper/2508.16465/) (2025)
- [Multi-Teacher Knowledge Distillation For Text Image Machine Translation](https://mlsanity.com/paper/2305.05226/) (2023)
- [Enhancing Hands in 3D Whole-Body Pose Estimation with Conditional Hands Modulator](https://mlsanity.com/paper/2603.14726/) (2026)
- [FuRPE: Learning Full-body Reconstruction from Part Experts](https://mlsanity.com/paper/2212.00731/) (2022)
- [PAFUSE: Part-based Diffusion for 3D Whole-Body Pose Estimation](https://mlsanity.com/paper/2407.10220/) (2024)
- ["Teaching Independent Parts Separately" (TIPSy-GAN) : Improving Accuracy and Stability in Unsupervised Adversarial 2D to 3D Pose Estimation](https://mlsanity.com/paper/2205.05980/) (2022)
- [A Coarse-to-Fine Human Pose Estimation Method based on Two-stage Distillation and Progressive Graph Neural Network](https://mlsanity.com/paper/2508.11212/) (2025)

## Abstract (arXiv)

We introduce DOPE, the first method to detect and estimate whole-body 3D human poses, including bodies, hands and faces, in the wild. Achieving this level of details is key for a number of applications that require understanding the interactions of the people with each other or with the environment. The main challenge is the lack of in-the-wild data with labeled whole-body 3D poses. In previous work, training data has been annotated or generated for simpler tasks focusing on bodies, hands or faces separately. In this work, we propose to take advantage of these datasets to train independent experts for each part, namely a body, a hand and a face expert, and distill their knowledge into a single deep network designed for whole-body 2D-3D pose detection. In practice, given a training image with partial or no annotation, each part expert detects its subset of keypoints in 2D and 3D and the resulting estimations are combined to obtain whole-body pseudo ground-truth poses. A distillation loss encourages the whole-body predictions to mimic the experts' outputs. Our results show that this approach significantly outperforms the same whole-body model trained without distillation while staying close to the performance of the experts. Importantly, DOPE is computationally less demanding than the ensemble of experts and can achieve real-time performance. Test code and models are available at https://europe.naverlabs.com/research/computer-vision/dope.