1 / 8

Coffee Talk: Non-Metric Recognition & Advantages of Metric Learning

This meta-analysis discusses the effectiveness of non-metric procedures in pattern recognition, specifically focusing on labeled faces, CalTech datasets, and SVM, highlighting the benefits of non-metric approaches over metric ones.

yates
Download Presentation

Coffee Talk: Non-Metric Recognition & Advantages of Metric Learning

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Good recognition is non-metric W.J. Scheirer, M.J. Wilber, M. Eckmann, T.E. Boult Pattern Recognition, August 2014, 47(2014)2721–2731 Coffee Talk

  2. Metric distances Meta-analysis on published results w.r.t. metric procedures Coffee Talk

  3. Labeled Faces in the Wild (LFW) performances Coffee Talk

  4. CalTech 101, 15 training images Coffee Talk

  5. CalTech 101, 30 training images Coffee Talk

  6. Multi-class SVM (non-metric) is better and faster than metric learning Coffee Talk

  7. Conclusions Many popular procedures are non-metric when applied pairwise Multi-class SVM One shot similarity Cosine similarity PLDA Tom-vs-Pete … it is unclear what advantage, if any, would be provided by enforcing the constraints of symmetry and the triangle inequality. …. What is the advantage of metric learning? Coffee Talk

More Related