A multiple instance dictionary learning method using functions of multipleinstances (DL-FUMI) is proposed to address target detection and two-classclassification problems with inaccurate training labels. Given inaccuratetraining labels, DL-FUMI learns a set of target dictionary atoms that describethe most distinctive and representative features of the true positive class aswell as a set of nontarget dictionary atoms that account for the sharedinformation found in both the positive and negative instances. Experimentalresults show that the estimated target dictionary atoms found by DL-FUMI aremore representative prototypes and identify better discriminative features ofthe true positive class than existing methods in the literature. DL-FUMI isshown to have significantly better performance on several target detection andclassification problems as compared to other multiple instance learning (MIL)dictionary learning algorithms on a variety of MIL problems.
展开▼