Research map / Video / Video generation
HumanMAC: Masked Motion Completion for Human Motion Prediction
Ling-Hao Chen, JiaWei Zhang, Yewen Li, Yiren Pang, Xiaobo Xia, Tongliang Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add DCT/iDCT representation (ablation) | HumanMAC · Human3.6M | ADE: 0.444 -> 0.369 | Improved |
| add DCT/iDCT representation (ablation) | HumanMAC · Human3.6M | APD: 7.191 -> 6.301 | Hurt |
| use Cosine scheduler instead of Sqrt (ablation) | HumanMAC · Human3.6M | ADE: 0.582 -> 0.369 | Improved |
| use Cosine scheduler instead of Linear (ablation) | HumanMAC · Human3.6M | ADE: 0.379 -> 0.369 | Improved |
| use 1000 noising and 100 DDIM steps instead of 100 and 10 (ablation) | HumanMAC · Human3.6M | ADE: 0.385 -> 0.369 | Improved |
| add skip connections in TransLinear (ablation) | HumanMAC · Human3.6M | ADE: 0.377 -> 0.369 | Improved |
| add skip connections in TransLinear (ablation) | HumanMAC · HumanEva-I | ADE: 0.208 -> 0.209 | No clear effect |
| increase DCT components L from 5 to 10 (ablation) | HumanMAC · HumanEva-I | ADE: 0.318 -> 0.209 | Improved |
| increase number of layers from 4 to 10 (ablation) | HumanMAC · HumanEva-I | APD: 6.554 -> 3.350 | Hurt |
| use masked completion HumanMAC (one loss, end-to-end) instead of multi-loss encoder-decoder methods | Human3.6M · Human3.6M | ADE: 0.389 -> 0.369 | Improved |
| use HumanMAC instead of baselines on diversity | Human3.6M · Human3.6M | APD: 6.301 vs 15.353 | Hurt |
Later papers that compared with this method
1 later accepted paper reports HumanMAC as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Learning Semantic Latent Directions for Accurate and Controllable Human Motion Prediction (ECCV 2024 · better on 2 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BeLFusion — BeLFusion: Latent Diffusion for Behavior-Driven Human Motion Prediction (ICCV 2023) · better on 2 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MMM: Generative Masked Motion Model (CVPR 2024 · similarity 0.88)
- MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation (AAAI 2024 · similarity 0.87)
- Seamless Human Motion Composition with Blended Positional Encodings (CVPR 2024 · similarity 0.86)
- SimpliHuMoN: Simplifying Human Motion Prediction (NeurIPS 2026 · similarity 0.86)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.86)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
