Gender recognition from frontal facial images within project FP7-ICT-247525-HUMAVIPS

SVM Adaboost - testing 2D data

<< zpět

Overview

The aim of this experiment is to test algorithm that learn Adaboost model based on linear classifiers learned by SVM (exactly, it will be used libOCAS as it was in experiment_02). For testing will be used artifically generated two dimensional dataset that can be easily plotted.

Training error and convex loss function should fall down (Monotony is not required).

Results


Figure 1. Iterations of algorithm

Figure 2. Training error and normalization constant

Figure 3. Convex loss function
<< zpět