Research map / 3D vision / Face and shape
Hi-LASSIE: High-Fidelity Articulated Shape and Skeleton Discovery From Sparse Image Ensemble
Chun-Han Yao, Wei-Chih Hung, Yuanzhen Li, Michael Rubinstein, Ming-Hsuan Yang, Varun Jampani
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 |
|---|---|---|---|
| Use Hi-LASSIE instead of LASSIE | articulated animal reconstruction · Horse, Cow, Sheep, Zebra, Tiger, Giraffe, Elephant, Kangaroo, Penguin | PCK@0.05: 58.0 -> 59.6; 62.4 -> 63.1; 55.5 -> 56.2; 63.3 -> 64.2; 62.4 -> 63.1; 60.5 -> 61.6; 40.3 -> 42.7; 31.5 -> 35.0; 40.6 -> | Improved |
| Use Hi-LASSIE instead of LASSIE | Pascal-Part animal reconstruction · Horse, Cow, Sheep | Overall IOU: 81.9 -> 83.4; 87.1 -> 88.1; 85.5 -> 86.3 | Improved |
| Use Hi-LASSIE instead of LASSIE | Pascal-Part animal reconstruction · Horse, Cow, Sheep | Part IOU: 38.2 -> 39.0; 35.1 -> 35.3; 43.7 -> 43.4 | Mixed |
| Use Hi-LASSIE instead of LASSIE | Pascal-Part animal reconstruction · Horse, Cow, Sheep | PCP: 78.5 -> 79.9; 77.0 -> 77.8; 74.3 -> 75.5 | Mixed |
| Add instance-specific part MLPs using frequency decomposition (ablation) | Hi-LASSIE · Pascal-Part and LASSIE animal ensembles | PCK@0.05: 56.8 -> 59.6; 57.7 -> 63.1; 53.6 -> 56.2; 59.7 -> 64.2; 63.0 -> 63.1; 59.9 -> 61.6; 40.8 -> 42.7; 31.5 -> 35.0; 41.2 -> | Mixed |
| Add instance-specific part MLPs using frequency decomposition (ablation) | Hi-LASSIE · Pascal-Part images: Horse, Cow, Sheep | Overall IOU: 80.4 -> 83.4; 80.0 -> 88.1; 79.5 -> 86.3 | Improved |
| Add feature MLPs (ablation) | Hi-LASSIE · Pascal-Part images: Horse, Cow, Sheep | Part IOU: 34.9 -> 39.0; 32.4 -> 35.3; 39.7 -> 43.4 | Improved |
| Add zoomed-in part silhouette loss L_part (ablation) | Hi-LASSIE · Pascal-Part images: Horse, Cow, Sheep | PCP: 79.2 -> 79.9; 77.4 -> 77.8; 72.0 -> 75.5 | Mixed |
| Add zoomed-in part silhouette loss L_part (ablation) | Hi-LASSIE · Pascal-Part images: Horse, Cow, Sheep | Overall IOU: 81.6 -> 83.4; 84.7 -> 88.1; 83.8 -> 86.3 | Mixed |
| Add feature MLPs (ablation) | Hi-LASSIE · Pascal-Part images: Horse, Cow, Sheep | PCP: 74.7 -> 79.9; 72.4 -> 77.8; 74.9 -> 75.5 | Mixed |
Later papers that compared with this method
2 later accepted papers report Hi-LASSIE as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- ARTIC3D: Learning Robust Articulated 3D Shapes from Noisy Web Image Collections (NeurIPS 2023 · better on 22 of 22 shared results)
- LEPARD: Learning Explicit Part Discovery for 3D Articulated Shape Reconstruction (NeurIPS 2023 · better on 13 of 13 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.
- LASSIE — LASSIE: Learning Articulated Shapes from Sparse Image Ensemble via 3D Part Discovery (NeurIPS 2022) · better on 17 of 18 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
- LASSIE: Learning Articulated Shapes from Sparse Image Ensemble via 3D Part Discovery (NeurIPS 2022 · similarity 0.91)
- ARTIC3D: Learning Robust Articulated 3D Shapes from Noisy Web Image Collections (NeurIPS 2023 · similarity 0.89)
- Hierarchical Kinematic Probability Distributions for 3D Human Shape and Pose Estimation From Images in the Wild (ICCV 2021 · similarity 0.88)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.88)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · similarity 0.87)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (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.
