neural-style
neural-style copied to clipboard
jcjohnson
Neural Algorithm to recognize artistic style
It's great project, thanks Johnson! Here we have a pictures repository where stored more than ten thousands artistic pictures (high resolutions) online, and people look for the artistic pictures via internet or mobile phone from time to time. Meanwhile, we are researching the feasibility to implement one recommendation mechanism based on the artistic style, that means, if the people is interesting at one specific artistic pictures, hence some other similar style of artistic pictures can be recommend to the people as well. Now the challenge to us is how to recognize the artistic style with different category for those ten thousands artistic picture with the Neural Algorithm. After checked the project of Neural-sytle (I didn’t study the detail yet), perhaps there is a chance to reuse some of the function as import from your project to achieve the purpose, but may I get your advice and how to move forward properly? thanks.
What sort of data do you have about each picture? Can they be [read: have you already] uniquely identified as representations of their real life counterparts?
Hi, thanks, Right now the pictures stored with JPG format only, and each picture was linked to the real life artist.
Can you give me an example of the data please? Or better yet, a link to the set? I am no expert, but I am quite interested in this problem.
Hi, thanks, you could check the website with example: http://www.baozhenart.com/category.php?id=72 Unfortunately it's Chinese version only, then you may need Google translator to understand.
And this is your website? Just need to check for legal reasons. Do you have access to the database?
On 8 July 2017 at 17:06, huapingchen [email protected] wrote:
Hi, thanks, you could check the website with example: http://www.baozhenart.com/category.php?id=72 Unfortunately it's Chinese version only, then you may need Google translator to understand.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/jcjohnson/neural-style/issues/416#issuecomment-313864981, or mute the thread https://github.com/notifications/unsubscribe-auth/ANWh-8MkN2FzXFBRrAK_UsSv3Zb5dX4Uks5sL6kYgaJpZM4OM3vs .
Doesn't Google Reverse Image Search already perform what you are trying to accomplish?
Hi, R-o-g, It's my friend own website, so there is no any leagl issue. Yes, we can access database, but the point here is not how to access the datatbase, my question is how to apply with the "artistic style of neurla algorthim". thanks anyway.
Thanks, I understand - but having access to the database if a prerequisite for doing the job easily.
I doubt there is a "quick" answer, but my inclination would have been to look into binary categorisation algorithms, at the same time as neural style.
I'm interested in this question, and if you pop me a note in a month or two, I might have my data scientists and engineers look at it for you, but only if we can share the data sets. ;-)
r
On 11 July 2017 at 07:18, huapingchen [email protected] wrote:
Hi, R-o-g, It's my friend own website, so there is no any leagl issue. Yes, we can access database, but the point here is not how to access the datatbase, my question is how to apply with the "artistic style of neurla algorthim". thanks anyway.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/jcjohnson/neural-style/issues/416#issuecomment-314342689, or mute the thread https://github.com/notifications/unsubscribe-auth/ANWh--VjvYts8J6aYRdDfstoDHpKKAXUks5sMxPQgaJpZM4OM3vs .