|
|
article-tgtimes-Synthetic-Ascii-Art.mw - tgtimes - The Gopher Times |
|
|
 |
git clone git://bitreich.org/tgtimes git://enlrupgkhuxnvlhsf6lc3fziv5h2hhfrinws65d7roiv6bfj7d652fid.onion/tgtimes (git://bitreich.org) |
|
|
 |
Log |
|
|
|
Files |
|
|
|
Refs |
|
|
|
Tags |
|
|
|
README |
|
|
|
--- |
|
|
|
article-tgtimes-Synthetic-Ascii-Art.mw (4099B) |
|
|
|
--- |
|
|
|
1 .SH tgtimes |
|
|
|
2 Synthetic ASCII Art |
|
|
|
3 .2C 57v |
|
|
|
4 . |
|
|
|
5 .PP |
|
|
|
6 When an entirely new way to solve problems is discovered, |
|
|
|
7 all sorts of medias, and not only the tech-oriented ones, |
|
|
|
8 are fond to publish abundantly about it. |
|
|
|
9 Be it quantum computing, blockchains, machine learning... |
|
|
|
10 Shortly after a new big toys like these comes-up, |
|
|
|
11 hackers come, and start experimenting with it, |
|
|
|
12 sometimes coming-up with entirely new way to use it. |
|
|
|
13 . |
|
|
|
14 .PP |
|
|
|
15 This time we are reviewing the combo of Machine Learning and ASCII art. |
|
|
|
16 . |
|
|
|
17 .PP |
|
|
|
18 I was expecting to present cute attempts at drawing images with computer-made text, but this is nothing of the sort. |
|
|
|
19 Be prepared to see Science at the service of Art. |
|
|
|
20 . |
|
|
|
21 .IP "Generated Typewriter Art" |
|
|
|
22 This research paper (no less!) shows that it is possible to write software for placing characters, |
|
|
|
23 later typed during 6 hours by a human operator (for this example). |
|
|
|
24 It is unsettling to see details much smaller than the characters themself be drawn on paper, |
|
|
|
25 along with shades of grey of various intensities. |
|
|
|
26 . |
|
|
|
27 .FS |
|
|
|
28 https://graphicsinterface.org/wp-content/uploads/gi2021-13.pdf |
|
|
|
29 .FE |
|
|
|
30 . |
|
|
|
31 .IP "Generated ASCII Art in 2010" |
|
|
|
32 This is possibly the state of the art of 2010 technology. |
|
|
|
33 It was announced in the yearly conference SIGGRAPH hence presented to an audience full of computer graphics engineers. |
|
|
|
34 The work of three researchers from Hong Kong, Xuemiao Xu, Linling Zhang and Tien-Tsin Wong, shows results of surprising accuracy. |
|
|
|
35 The story does not tell whether there ever was a job offer "looking for ASCII artists for a scientific experiment" posted on the job board of the Chinese University of Hong Kong. |
|
|
|
36 While the paper contains the complete math used, it also illustrates and |
|
|
|
37 explains methods to achieve this level of accuracy. |
|
|
|
38 And no, it is not exactly machine learning, but hand-crafted strategies, combined statistics and other data massaging. |
|
|
|
39 After all, it was published five years before things like Tensor Flow were introduced... |
|
|
|
40 . |
|
|
|
41 .FS |
|
|
|
42 http://www.cse.cuhk.edu.hk/~ttwong/papers/asciiart/asciiart.html |
|
|
|
43 .FE |
|
|
|
44 . |
|
|
|
45 .IP "Generated ASCII Art in 2017" |
|
|
|
46 Is seven years enough time to improve upon that previous achievement? |
|
|
|
47 Quoting the previous paper as well as others in its own work, |
|
|
|
48 Osamu Akiyama of the Osaka Faculty of Medicine kept the ball rolling. |
|
|
|
49 This throws the big guns of machine learning to reach higher skies. |
|
|
|
50 Its input data were Japaneses BBS such as 5chan (2chan) or Shitaraba, |
|
|
|
51 which extends the ASCII set to all of unicode, notably the CJK set. |
|
|
|
52 If the result of the paper are not enough to convince you, |
|
|
|
53 the "Bad Apple" often used as a video demo in the Asian market have been converted in its entirety. |
|
|
|
54 Something out of reach if doing every frame by hand. |
|
|
|
55 The TensorFlow and Python code used is released publicly, |
|
|
|
56 and an online demo is offered for the curious. |
|
|
|
57 . |
|
|
|
58 .FS |
|
|
|
59 https://nips2017creativity.github.io/doc/ASCII_Art_Synthesis.pdf |
|
|
|
60 .FE |
|
|
|
61 .FS |
|
|
|
62 https://nips2017creativity.github.io/ |
|
|
|
63 .FE |
|
|
|
64 .FS |
|
|
|
65 https://yewtu.be/watch?v=8GulN69Cgbg |
|
|
|
66 .FE |
|
|
|
67 .FS |
|
|
|
68 https://www.vice.com/en/article/zmymwx/machine-learning-ascii-art-neural-net |
|
|
|
69 .FE |
|
|
|
70 .FS |
|
|
|
71 https://github.com/OsciiArt/DeepAA |
|
|
|
72 .FE |
|
|
|
73 . |
|
|
|
74 .PP |
|
|
|
75 Is it so futile? Not so sure. |
|
|
|
76 After all, representing anything with a computer is a matter of making a reality fit onto something terribly awkward and unnatural: a display. |
|
|
|
77 The pixels, the square elements praised for providing a grid to throw data at, are promising, but themself have their quirks to be worked around. |
|
|
|
78 For instance, sub-pixel geometry uses the same techniques as those presented by these papers for improving the realism of images beyond what a single pixel can offer. |
|
|
|
79 It is, for ASCII art like for anything else, a matter of representing something, real or fictious, through a medium of some kind. |
|
|
|
80 . |
|
|
|
81 .PP |
|
|
|
82 ASCII art has the ability to fit an image somewhere where there could only be text. |
|
|
|
83 For the example of a train station concourse with a large split-flap display: |
|
|
|
84 for displaying a big arrow at the end of the service, |
|
|
|
85 replacing the display by an equally large color screen can be costly and much more power-hungry, |
|
|
|
86 while an ASCII arrow on that existing display would be consuming no power for that still image. |
|