1 / 10

REU Report Meeting Week 4

REU Report Meeting Week 4. Patrick Hynes. Topic. One shot recognition Extending to Action recognition. Animals with Attributes. Caltech 101 Experiments. Goal, compare results to Fei-fei,Fergus , Perona All 101 Categories Training set of 3,6,15 examples per class

shae
Télécharger la présentation

REU Report Meeting Week 4

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. REU Report Meeting Week 4 Patrick Hynes

  2. Topic • One shot recognition • Extending to Action recognition

  3. Animals with Attributes

  4. Caltech 101 Experiments • Goal, compare results to Fei-fei,Fergus, Perona • All 101 Categories • Training set of 3,6,15 examples per class • Testing set of 50 examples per class

  5. Caltech 101 Images

  6. Some Results • Working with Antonio to compare our results • His code is using the same examples repeatedly, mine is choosing examples from each category randomly

  7. Some Results • Optimization Function

  8. Some Results • Using PCA with 3 training examples • 3 training examples over 300 optimization iterations, 1 microset 8.21% mean accuracy • Using PCA over training and testing examples • 3 examples 1000 iter. 1microset 8.52% mean accuracy • 3 examples 600 iter. 10 microsets 9.28% • 6 examples 300 iter. 1 microset 10.37%

  9. Areas to test and improve • Experimenting with • different PCA data • number of microsets • Proper number of iterations for optimization • Increased one-shot repetitions in checking accuracy • Averaging results from various data

More Related