Development of a deep learning model for crowd counting

How to use this code

1. for Training:

The training part is an updated an adjusted version of the Crowd Counting Framework by Gao et al. (https://github.com/gjy3035/C-3-Framework).

• requirements: Python 3.X, Pytorch 2.2, other libs in requirements.txt
• set variables in config.py: dataset (SHHA, SHHB or SHHM), exp path, training with or without GPU support, other hyperparameters
• in datasets/SHHX/setting.py set path to data with labels
• run train.py

2. to use the crowd counting application:

• additional required libraries: PIL and customtkinter
• change path to 4 models that will be compared in gui.py
• start gui.py, select image in file system and see results