Curriculum for Crowd Counting -- Is it Worthy?
This addresses the debate over curriculum learning's effectiveness for supervised learning in computer vision, specifically for crowd counting, but is incremental as it applies an existing technique to a domain.
The paper investigates the impact of curriculum learning on crowd counting using density estimation, finding that it improves model performance and reduces convergence time based on 112 experiments with six CL settings and eight models.
Recent advances in deep learning techniques have achieved remarkable performance in several computer vision problems. A notably intuitive technique called Curriculum Learning (CL) has been introduced recently for training deep learning models. Surprisingly, curriculum learning achieves significantly improved results in some tasks but marginal or no improvement in others. Hence, there is still a debate about its adoption as a standard method to train supervised learning models. In this work, we investigate the impact of curriculum learning in crowd counting using the density estimation method. We performed detailed investigations by conducting 112 experiments using six different CL settings using eight different crowd models. Our experiments show that curriculum learning improves the model learning performance and shortens the convergence time.