视频/图像损伤修复是一个比较综合的范畴,包括去伪影/去块/去噪(artifacts reduction, deblocking, denoising)。为帮助选手对赛题快速理解,这里借用了Artifacts Reduction Convolutional Neural Network供选手参考。
This is a Tensorflow implementation of Artifacts Reduction Convolutional Neural Network.
-
Download BSDS500 dataset, rename the dataset folder as
BSDS500
and put it in./data
directory -
Download LIVE1 dataset, rename the dataset folder as
databaserelease2
and put it in./data
directory -
change matlab working directory to
the/repository/path/data/code/source
-
run
extract_data
on matlab console
-
cd src
-
python train.py
-
cd src
-
python test.py