/CSRNet_pytorch

personal implementation of csrnet which is designed for crowd counting

Primary LanguageJupyter Notebook

CSRNet_pytorch

personal implementation of csrnet which is designed for crowd counting

before use the DatasetConstructor.py for importing and preprocessing the dataset, need to transform the gt from .mat into density map(.npy) and saved in the disk first.(I use the Shanghaitech dataset.) which means run "generate_density_map first"