A New High Performance Approach for Crowd Counting Using Human Filter

One of the tasks of the crowd monitoring system is to estimate the number of people in the crowd and issue a warning when it exceeds the allowed threshold. Previous approaches often used multi-column CNN to estimate density maps and thereby estimate the count. However, the amount of information learned from crowd datasets is very small. On the other hand, the confusion between people and other objects such as buildings, trees, rocks, etc (background noise) will affect the density map estimation. In this paper, we focus to solve these two problems and propose a model called Counting using Human Filter (CHF) which consists of two modules: The first one is a feature extractor from a crowd image based on the VGG-16 model to estimate the density map. This module will take advantage of the features learned from the ImageNet dataset. The second one is the human filter used to weight each pixel of the density map. Two modules are combined by element-wise multiplication. We evaluate the estimated results of the model with MAE, MSE metric and assess the quality of density maps according to PSNR, SSIM. Experiments show that our approach estimates the number of people better than the previous methods when evaluating on the ShanghaiTech, UCF_CC_50, UCF-QRNF datasets. Regarding the complexity of the model, our method shares parameters between two modules so it halved the number of parameters compared to previous methods such as Switch-CNN, SSC, ADCrowdNet.