Data and codes for "SCANet: Self-paced semi-curricular attention network for non-homogeneous image dehazing"

This record contains the data and codes for the paper "SCANet: Self-Paced Semi-Curricular Attention Network for Non-Homogeneous Image Dehazing" published in 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). RequirementPython 3.7Pytorch 1.9.1Network ArchitectureTrainPlace the training and test image pairs in the data folder.Run data/makedataset.py to generate the NH-Haze20-21-23.h5 file.Run train.py to start training.TestPlace the pre-training weight in the checkpoint folder.Place test hazy images in the input folder.Modify the weight name in the test.py.parser.add_argument("--model_name", type=str, default='Gmodel_40', help='model name')Run test.pyThe results is saved in output folder.Pre-training Weight DownloadThe weight40 Gmodel_40.tar for the NTIRE2023 val/test datasets, i.e., the weight used in the NTIRE2023 challenge.The weight105 Gmodel_105.tar for the NTIRE2020/2021/2023 datasets.The weight120 Gmodel_120.tar for the NTIRE2020/2021/2023 datasets (Add the 15 tested images as the training dataset).

本记录包含了发表于2023年IEEE/CVF计算机视觉与模式识别会议（CVPRW）工作坊的论文《SCANet：针对异构图像去雾的自适应速度半课程注意力网络》（SCANet: Self-Paced Semi-Curricular Attention Network for Non-Homogeneous Image Dehazing）所涉及的数据和代码。技术要求：Python 3.7，Pytorch 1.9.1。网络架构：将训练和测试图像对放置于数据文件夹中。运行data/makedataset.py脚本以生成NH-Haze20-21-23.h5文件。执行train.py脚本以启动训练过程。测试：将预训练权重放置于检查点文件夹中，将测试模糊图像放置于输入文件夹中。在test.py.parser.add_argument函数中修改权重名称，参数为--model_name，类型为字符串，默认值为'Gmodel_40'，帮助信息为'model name'。运行test.py脚本。结果将保存在输出文件夹中。预训练权重下载：下载40 Gmodel_40.tar权重，用于NTIRE2023 val/test数据集，即NTIRE2023挑战赛中所使用的权重。下载105 Gmodel_105.tar和120 Gmodel_120.tar权重，分别对应NTIRE2020/2021/2023数据集（将15张测试图像添加为训练数据集）。