Open source demosaicing artifact to repair damaged comics without stress

In recent years, with the rise of artificial intelligence, AI tools for image restoration have become more and more numerous. Recently, deeppomf released a DeepCreamPy project on GitHub, which can help you make up for the mosaic images in H-manga.
Please add image description
After the software was uploaded and shared, it was downloaded more than 10,000 times within a week. At present, the project has obtained 3965 Stars and 306 Forks on GitHub (project address: https://github.com/deeppomf/DeepCreamPy)
Please add image description
However, the limitations of the current software are still very large, and only some simple fixes can be completed. For example, in this article, after drawing a green line on the original image at any position for coding, it can be restored through the DeepCreamPy software.

In order to make this software achieve better results, the author collected more than 100,000 uncoded original images in just a few months, and he also released many comparison images.

Instructions

64-bit Windows users can directly download the exe program (download address: https://github.com/deeppomf/DeepCreamPy/releases/latest). If you are not using a Windows system, you need to compile it yourself. If you need to retrain the model, The following tools are required:

● Python 3.6

● TensorFlow 1.10

Loud 2.2.4

● Pillow

● h5py

Tensorflow, Keras, Pillow, and h5py can all be installed with:

$ pip install -r requirements.txt

If you want to see more partners, you can go to the project details page to check.

More Python videos, source code, and data can be obtained for free in the group 660193417

Guess you like

Origin blog.csdn.net/m0_67575344/article/details/124203624