Apparently there was an explosion on fanart usage, so I’m adding fuel to the fire. For the record, I passed up wildarmshero’s post, and pretty much every other related post. There was simple generation of interest, which led to reading some related posts.
OFP (link)
After reading the statement, I have to identify a bit with omomomomo. It’s silly because it tends to give off the impression, GTFOG. In a perfect function of this OFP, everyone would contact the original artist and request permission to post the art along with names, links, etc. Fantastic, but what a pain in the ass, likely involving waiting.
If artists don’t want their art shared [by foreigners or whoever], then use the OFP badge or state their exact conditions, preferably with the notice in at least one Western language.
Reading List
- Mistakes of Youth (wah, artist/blogger)
- 見ないで！ひとり言 (Hinano, artist/blogger)
- [did read others via Continuing World]
Best point made: go to Pixiv
On an unrelated note, I didn’t even know I had a pixiv account… wtf
2nd best point made: accredit what you can (artist name, link, homepage, etc)
Re-enforce Pixiv
Image blogs are a popular item, and there are numerous Japanese posters who re-blog other artists works [here's one I like recently, for the blend]. Having subscribed to a few of these, I think one of the OFP’s bullets is regularly breached; contacting the artists for permission, but whatever.
The most reassuring point that comes from seeing posts on these blogs, is that most anime/manga-esque works come from pixiv, and one can go right to the pixiv page. I believe it is a good situation for both parties.
In the spirit of accreditation, here’s two arts I recently liked: one brushing teeth and one Ms. Langley. (Images not featured :P)
Tech Thought
This issue that was raeged about is not simply going to be solved, due to unawareness. Now, if we had something like MusicBrainz acoustic fingerprinting, for images (or partial images), open registration/submission for identifiers and artists, we wouldn’t have a problem of knowning who really created what we see. Probably could use Eigen pairs, though I have no experience with that for color images ^_^
What I didn’t like about this was that it all seemed like a sudden epiphany that blew up like a ragebomb9001, and that suddenly everyone was concerned. This is by far not a new problem, after all, but yeah, eventually it was just the same old drama train going on. In the end I was fairly amused by seeing the usual culprits raging and trolling about.
I secretly hoped that there would be a couple of influential retards who would call for a blogosphere-wide witch hunt to name and shame those who didn’t follow this mere unspoken rule - as it were- which would create a massive shitstorm. But I guess that kind of drama will occur some other time. Bad for my schadenfreude.
ROFL nekosasu. I’m not sure why this even came up (what made whoever post on it). Whatever it was, I think everyone was just doing things naturally, based on their level of awareness. If someone is getting images from google image search, it’s likely that there’s not going to be proper info for mentions.
I don’t think anyone intends to harm or cause misfortune by not crediting art [or not knowing to do so].
what JP was saying about searching images as if they were text, you can do that with music, in certain tech-focused libraries. “Music” is misleading - it’s just syntax. I suppose it’s harder with images.
on another point - with music, as far as I’ve seen, if people come upon a song and the artist is unknown, there’s almost an instantaneous blitz to find out who composed it. This is similar to the whole SAUCE?!!??! meme, though with SAUCE it’s more about finding the source anime, by which further doujin is found. It doesn’t really work that way with music, since there’s not nearly as much doujin music.
also, is this a new blog theme? I dig it!
@lelangir I wouldn’t say it’s harder with images, but it’s more difficult to be precise, especially on a grand scale. Face recognition works based on Eigen pairs (basically a big matrix), and what happens is that each pair aligns to give a certain value. If the values match within a certain range, the image is considered similar enough.
Searching images with text might be tougher, since even now, research is in full-force for object detection, but it’s all based on training these AI algorithms. The most efficient method would definitely be Eigen vector-value matching, but that goes with the assumptions 1) the image is identified in the databased somewhere 2) you have an image you wish to compare (I suppose even sketches would work, tracing definitely). After the matching is done, the result may still be partial or hazy, but close.
That would be some massive database, since I’m pretty sure there is more artwork in the world than there are musical pieces… but that’s arguable.
[...] is a continuation post on that fanart citation circlejerk, but a tangent that began here (scroll [...]