{"id":145,"date":"2025-12-12T00:39:11","date_gmt":"2025-12-12T00:39:11","guid":{"rendered":"https:\/\/mscvprojects.ri.cmu.edu\/2025team13\/?page_id=145"},"modified":"2025-12-13T03:27:31","modified_gmt":"2025-12-13T03:27:31","slug":"results","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/2025team13\/results\/","title":{"rendered":"\u00a0Results & Conclusion"},"content":{"rendered":"\n

Results & Conclusion<\/strong><\/h1>\n\n\n\n

Results<\/strong><\/h2>\n\n\n\n

Bridging the Sim2Real Gap with Generative Models<\/strong><\/h3>\n\n\n\n

Given that our point-map regressor uses a frozen SD-DINO encoder as its backbone, one plausible explanation for its ability to generalize to real humans is that the encoder alone provides strong domain robustness. To assess the importance of the synthetic data generation pipeline proposed, we compare two point-map regressors evaluated on RGB images of humans from the MoVi dataset. The first model is trained solely on robot-rendered images, while the second is trained on our synthetically generated human images described in our data generation pipeline.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

As illustrated in the figure above, our results reveal a clear contrast:<\/p>\n\n\n\n

the regressor trained on robot imagery fails to generalize to real humans, whereas the regressor trained on synthetic human images achieves substantially better performance. This indicates that the synthetic data generation pipeline is essential for bridging the domain gap and enabling effective transfer to real human imagery.<\/p>\n\n\n\n

Generalization to Diverse Body Proportions<\/strong><\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

We assess our model\u2019s ability to handle extreme body proportions by evaluating it on an unusually tall basketball player whose limb lengths and overall stature lie far outside the distribution of typical training subjects, as demonstrated in the figure above. Despite this substantial deviation from canonical human shape, the model continues to produce stable, multi-layer 3D point correspondences for every pixel. Notably, these predicted correspondences remain metrically consistent and align well to a fixed-size humanoid robot without requiring explicit re-scaling or additional optimization. This indicates that our feedforward pipeline effectively bridges the embodiment gap even when confronted with highly stretched limb ratios. While this case study examines a single extreme example, it highlights the model\u2019s ability to extract shape-invariant geometric cues, suggesting promising generalization to a broader range of atypical body types.<\/p>\n\n\n\n

In the Wild Example<\/strong><\/p>\n\n\n\n