Skip to content

This is an unofficial implement of the arXiv paper Dual Path Multi-Scale Fusion Networks with Attention for Crowd Counting by PyTorch.

Notifications You must be signed in to change notification settings

pxq0312/SFANet-crowd-counting

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SFANet-crowd-counting

This is an unofficial implement of the arXiv paper Dual Path Multi-Scale Fusion Networks with Attention for Crowd Counting by PyTorch.

Prerequisite

Python 3.7

Pytorch 1.1.0

Code structure

density_map.py To generate the density map and attention map.

dataset.py and transforms.py For data preprocess and augmentation.

models.py The structure of the network.

train.py To train the model.

eval.py To test the model.

Train & Test

For training, run

python train.py --dataset="SHA" --data_path="path to dataset" --save_path="path to save checkpoint"

For testing, run

python eval.py --dataset="SHA" --data_path="path to dataset" --save_path="path to checkpoint"

Result

ShanghaiTech part A: epoch367 MAE 60.43 MSE 98.24

ShanghaiTech part B: epoch432 MAE 6.38 MSE 10.99

About

This is an unofficial implement of the arXiv paper Dual Path Multi-Scale Fusion Networks with Attention for Crowd Counting by PyTorch.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages