search

bupt-ai-cz / LLVIP

LLVIP: A Visible-infrared Paired Dataset for Low-light Vision
623 stars 64 forks source link

questions about image fusion part experiments #16

Closed challow0 closed 2 years ago

challow0 commented 2 years ago

Thanks for your great work ~ I have some questions about the experiments. I want to know how many images you used in the image fusion part experiment. Did you use all the images for training the FusionGAN and Densefuse algorithm? Or just the images in the directory? And what are the parameters you used? Looking forward to your reply.

ShenXinchang commented 2 years ago

Thanks for your great work ~ I have some questions about the experiments. I want to know how many images you used in the image fusion part experiment. Did you use all the images for training the FusionGAN and Densefuse algorithm? Or just the images in the directory? And what are the parameters you used? Looking forward to your reply.

Hi~ We used a subset of LLVIP dataset in all image fusion experiments. The subset has 1000+ image pairs and includes all the scenes in LLVIP dataset. The training parameters for FusionGAN and Densefuse experiments are defult setting. Due to the impact of Covid-19, we will release the training data and the subset after a short period of time. Best wishes.

challow0 commented 2 years ago

@ShenXinchang Thanks for your reply. Wish you can release it soon.😄