Weight-loss control sampling for the training of boosted pedestrian detectors

When we apply AdaBoost in pedestrian detection, a large number of examples are needed to train a detector. Except for designing features, a reasonable utilization of training examples is also significant to the detection accuracy and training time. In this paper, we propose a new method, named Weight-Loss Control Sampling (WLCS), to deal with the negative training examples by improving the training process of AdaBoost. The WLCS updates the negative training set by sampling hard examples from original negative set. It determines when to implement sampling process by the weight-loss of the negative training set. And we reduce the capacity of negatives after each sampling to accelerate the training. In this paper, we implement experiments to prove that the WLCS can train a better classifier in a short training time via a specific pedestrian detection method.