Evaluation-of-Neural-Network-Image-Dehazing-Methods

This is the final year project of Weijia Sun.

result_selfmade: This file contains the results using the self-made testing dataset under the epoch 30.

real-dehazed: This is the file containing the dehazed image using the pre-trained model.

checkpoints_new: This is the file generated in the training process, containing comparison between different epochs.

hazy: This file contains the compressed images after adding haze to the haze-free SYNS dataset, which is the data used in the experiment.

hazy beta=30: This file contains the original images after adding haze to the haze-free SYNS dataset. In the adding haze process, the beta is set to be 30, which is not used in the experiment.

hazy_beta=10_compressed: This file contains the original images after adding haze to the haze-free SYNS dataset. In the adding haze process, the beta is set to be 10, which is used in the experiment.

outrange: This file contains the transmission map of images from SYNS dataset.

png: This file contains the PNG format data, which are converted from HDR data format int he SYNS website.