FGENet: Fine-Grained Extraction Network for Congested Crowd Counting

TL;DR

FGENet is an end-to-end crowd counting model that directly localizes individuals with high accuracy, employing a fusion module and a robust loss function to outperform existing methods, especially in high-density scenarios.

Contribution

The paper introduces FGENet, a novel model that directly predicts individual coordinates, utilizing a new fusion module and loss function to improve accuracy over density map-based methods.

Findings

Achieves 3.14 MAE improvement on ShanghaiTech Part A dataset.

Surpasses previous benchmarks with 30.16 MAE improvement on UCF_CC_50.

Demonstrates robustness in high-density crowd images.

Abstract

Crowd counting has gained significant popularity due to its practical applications. However, mainstream counting methods ignore precise individual localization and suffer from annotation noise because of counting from estimating density maps. Additionally, they also struggle with high-density images.To address these issues, we propose an end-to-end model called Fine-Grained Extraction Network (FGENet). Different from methods estimating density maps, FGENet directly learns the original coordinate points that represent the precise localization of individuals.This study designs a fusion module, named Fine-Grained Feature Pyramid(FGFP), that is used to fuse feature maps extracted by the backbone of FGENet. The fused features are then passed to both regression and classification heads, where the former provides predicted point coordinates for a given image, and the latter determines the confidence level for each predicted point being an individual. At the end, FGENet establishes correspondences between prediction points and ground truth points by employing the Hungarian algorithm. For training FGENet, we design a robust loss function, named Three-Task Combination (TTC), to mitigate the impact of annotation noise. Extensive experiments are conducted on four widely used crowd counting datasets. Experimental results demonstrate the effectiveness of FGENet. Notably, our method achieves a remarkable improvement of 3.14 points in Mean Absolute Error (MAE) on the ShanghaiTech Part A dataset, showcasing its superiority over the existing state-of-the-art methods. Even more impressively, FGENet surpasses previous benchmarks on the UCF\_CC\_50 dataset with an astounding enhancement of 30.16 points in MAE.