1 / 6

Robust Face Detection using Random Projections

Random projections randomly project high-dimensional face data into a lower dimension for classification. The theory ensures robustness. Project from 50,000 to 1000 dimensions preserving distances. Test datasets include Essex, Sheffield, Georgia Tech, and Caltech Faces. Results show SVM classification with 90% training and 10% testing for various face datasets. Planned work includes exploring different probability distributions and kernel methods on the projected data. Random Projections are promising for face recognition but sensitive to background changes compared to PCA.

yepa
Télécharger la présentation

Robust Face Detection using Random Projections

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. Face detection using Random projections Sunil Khanal

  2. Random Projections Randomly project high dimensional data into low dimension Given , project it to using projection matrix     : Can use               to generate each element of the projection matrix Use the projection matrix to reduce data to 100-1000 dimensions Use the output for classification

  3. The theory Robustness (  ) of a concept class: Letberandomprojections of        . Given a threshold   , and target dimension k Projectingfrom a 50,000 dimensionspaceto 1000 dimension preserves pairwisedistancestowith          % probability

  4. Datasets tested Essex Faces (Expression variations, 360 images) Sheffield Faces (Pose variation, 575 images) Georgia Tech Faces (Similar lighting condition and background, 750 images) Caltech Faces (Varying lighting condition and background, 436 images)

  5. Results • - SVM classification with polynomial kernel (degree 3) • 90% training, 10% testing Sheffield (20 persons) Georgia Tech (50 persons) Essex (18 persons) Caltech (26 persons)

  6. Planned Work Interpreting the projection Explore different probability distributions for calculating the projection matrix Explore probabilistic kernel methods (esp. Gaussian product kernel) on the projected data RP seemingly works well for faces, but sensitive to background changes Comparison against PCA/feature based approaches

More Related