Timezone: »
An algorithm for learning fast multiclass object detection cascades is introduced. It produces multiresolution (MRes) cascades, whose early stages are binary target vs. nontarget detectors that eliminate false positives, late stages multiclass classifiers that finely discriminate target classes, and middle stages have intermediate numbers of classes, determined in a datadriven manner. This MRes structure is achieved with a new structurally biased boosting algorithm (SBBoost). SBBost extends previous multiclass boosting approaches, whose boosting mechanisms are shown to implement two complementary datadriven biases: 1) the standard bias towards examples difficult to classify, and 2) a bias towards difficult classes. It is shown that structural biases can be implemented by generalizing this classbased bias, so as to encourage the desired MRes structure. This is accomplished through a generalized definition of multiclass margin, which includes a set of bias parameters. SBBoost is a boosting algorithm for maximization of this margin. It can also be interpreted as standard multiclass boosting algorithm augmented with margin thresholds or a costsensitive boosting algorithm with costs defined by the bias parameters. A stage adaptive bias policy is then introduced to determine bias parameters in a data driven manner. This is shown to produce MRes cascades that have high detection rate and are computationally efficient. Experiments on multiclass object detection show improved performance over previous solutions.
Author Information
Ehsan Saberian (Netflix)
Nuno Vasconcelos (UC San Diego)
More from the Same Authors

2016 Poster: Large Margin Discriminant Dimensionality Reduction in Prediction Space »
Ehsan Saberian · Jose Costa Pereira · Nuno Nvasconcelos · Can Xu 
2016 Poster: Object based Scene Representations using Fisher Scores of Local Subspace Projections »
Mandar D Dixit · Nuno Vasconcelos 
2012 Poster: On the connections between saliency and tracking »
Vijay Mahadevan · Nuno Vasconcelos 
2012 Poster: Recognizing Activities by Attribute Dynamics »
Weixin Li · Nuno Vasconcelos 
2011 Poster: Maximum Covariance Unfolding : Manifold Learning for Bimodal Data »
Vijay Mahadevan · Chi Wah Wong · Jose Costa Pereira · Tom Liu · Nuno Vasconcelos · Lawrence Saul 
2011 Poster: Multiclass Boosting: Theory and Algorithms »
Mohammad J Saberian · Nuno Vasconcelos 
2010 Spotlight: A biologically plausible network for the computation of orientation dominance »
Kritika Muralidharan · Nuno Vasconcelos 
2010 Poster: Variable margin losses for classifier design »
Hamed MasnadiShirazi · Nuno Vasconcelos 
2010 Poster: A biologically plausible network for the computation of orientation dominance »
Kritika Muralidharan · Nuno Vasconcelos 
2010 Poster: Boosting Classifier Cascades »
Mohammad J Saberian · Nuno Vasconcelos 
2008 Poster: On the Design of Loss Functions for Classification: theory, robustness to outliers, and SavageBoost »
Hamed MasnadiShirazi · Nuno Vasconcelos 
2008 Spotlight: On the Design of Loss Functions for Classification: theory, robustness to outliers, and SavageBoost »
Hamed MasnadiShirazi · Nuno Vasconcelos 
2007 Poster: The discriminant centersurround hypothesis for bottomup saliency »
Dashan Gao · Vijay Mahadevan · Nuno Vasconcelos