Object Correspondence as a Machine Learning Problem
Abstract
We propose machine learning methods for the estimation of\ndeformation fields that transform two given objects into each other, thereby establishing a dense point to point correspondence. The fields are computed using a modified support vector machine\ncontaining a penalty enforcing that points of one object\nwill be mapped to ``similar‘‘ points on the other one. Our system,\nwhich contains little engineering or domain knowledge, delivers\nstate of the art performance. We present application results including close to\nphotorealistic morphs of 3D head models.
Cite
Text
Schölkopf et al. "Object Correspondence as a Machine Learning Problem." International Conference on Machine Learning, 2005. doi:10.1145/1102351.1102449Markdown
[Schölkopf et al. "Object Correspondence as a Machine Learning Problem." International Conference on Machine Learning, 2005.](https://mlanthology.org/icml/2005/scholkopf2005icml-object/) doi:10.1145/1102351.1102449BibTeX
@inproceedings{scholkopf2005icml-object,
title = {{Object Correspondence as a Machine Learning Problem}},
author = {Schölkopf, Bernhard and Steinke, Florian and Blanz, Volker},
booktitle = {International Conference on Machine Learning},
year = {2005},
pages = {776-783},
doi = {10.1145/1102351.1102449},
url = {https://mlanthology.org/icml/2005/scholkopf2005icml-object/}
}