quantombone / exemplarsvm

Ensemble of Exemplar-SVMs for Object Detection and Beyond
http://www.cs.cmu.edu/~tmalisie/projects/iccv11/index.html
MIT License
430 stars 155 forks source link

training should allow for max image dim #53

Open quantombone opened 12 years ago

quantombone commented 12 years ago

This way, if we give the system very large images training time is still bounded.

quantombone commented 12 years ago

This is more important that I previously thought...