While thinking that it is not the content to write in Qiita.
For those who do not know programming, we will explain how to install the Deep Learning AA Auto Tracer (DeepAA). For AA clusters and Yaruo clusters. Click here for details on DeepAA ↓ Making ASCII art with deep learning
If you have Windows 7 or later, it's probably okay.
Anaconda installer archive https://repo.continuum.io/archive/ Click on ʻAnaconda3-4.2.0-Windows-x86_64.exe` to download from. After downloading, run it and install according to the installer.
Open a command prompt (search for windows and type "cmd" in the child to get it). Execute the following command (enter the content after the "$" on each line).
$ pip install keras==1.2.0
$ pip install tensorflow
Download the installer from the official PyCharm website. https://www.jetbrains.com/pycharm/ After downloading, install according to the installer.
github repository
https://github.com/OsciiArt/DeepAA
Then click "Clone or Download" and then "Download Zip" to download the DeepAA zip.
After downloading, unzip it.
Place the unzipped folder DeepAA-master
in any folder you like.
Start Pycharm. You will be asked a lot at the first startup, but let's click yes for all.
Click [File]-[Open] from the menu on the upper right.
Select the saved "Deep AA-master".
Select deepaa_output.py
from the left panel and double-click it.
Right-click on the screen that opens to the right and click Run'deepaa_output'
on the opened panel.
Execution is successful when AA output starts at the bottom of the screen.
Here
Download ʻAA Convert.7z from. After downloading, unzip the file. Execute ʻAsciiArtConverter08.exe
inside
(Although it is AA automatic conversion software, I will borrow only the line art creation function.)
Click [File]-[Open Image File] from the menu on the upper left.
Select the prepared image you want to convert.
Click Contour Extraction Preview
to draw a line art. Adjust other parameters on the panel for better line art extraction.
After extracting the line art, right-click on the line art and click Save outline image
on the opened panel.
Select "Deep AA-master" as the save destination folder and save.
Set the contents of deepaa_output.py in PyCharm.
Rewrite the contents of # parameter
as follows.
17th line
image_path = 'test_image.png' #Specify the image to convert
Changed test_image.png
to the file name of the file saved earlier.
18th line
image_width = 550 #Image size conversion settings.Enter the converted width. ''If not converted. 1st setting: 550
Adjust the size by changing the 550
of. The value corresponds to the width of the adjusted image.
20th line
multi_output = False #Output 18 pixels by shifting each pixel vertically
If you change False
to True
, the image will be shifted by 1 pixel and output 18 times at once (optional).
Right-click and click Run'deepaa_output'
on the open panel. If AA is output as expected, it is successful.