Averaging Forest for Online Vision
Hassab Elgawi Osman · Journal of Advanced Computational Intelligence and Intelligent Informatics · 2009
In this study we consider vision as a binary classification problem, where an ensemble of decision-tree-based classifiers is trained on-line, new images are continuously added and the recognition decision is made without delay. Ensemble of decision trees is combined into aforestclassifier usingaveraging, generate an on-lineRandom Forest(RF) classifier. First we employ object descriptor model based on a bag of covariance matrices, to represent an object features, then run our on-line RF learner to select object descriptors and to learn object classifiers. Validation of our proposal with empirical studies in the GRAZ02 dataset domain demonstrates its superior performance over histogram-based counterparts, yielding object recognition performance comparable to state-of-the-art standard RF, AdaBoost, and SVM classifiers, even when only 10% of the training examples are used.