Generative Landmarks

The proposed method is capable of learning landmarks for various body regions, including faces and hands. The training set consists of roughly 18k frames (about 10 minutes each) for each domain – marked and unmarked, scaled down to resolution 128 × 128.

Data and Resources

Cite this as

D. Ferman, G. Bharaj (2024). Dataset: Generative Landmarks. https://doi.org/10.57702/zn7km0q5

DOI retrieved: December 16, 2024

Additional Info

Field Value
Created December 16, 2024
Last update December 16, 2024
Defined In https://doi.org/10.48550/arXiv.2104.04055
Author D. Ferman
More Authors
G. Bharaj