3D porn sort and duplicate search program

>>4686713
I just use VisiPics. It's free and gets the job done in a less complicated and more automated way.
I just removed 5k images (2~GB) yesterday.

File: 3DFileViewer_2017-07-13_22-36-10.png (21KB, 1027x737px) Image search: [Google]

21KB, 1027x737px

>>4686736

... well fuck then, thanks for the tip. Guess I still learned stuff making it.

for instance, hentai color tends to fit a line of best fit

>>4686753
Haha, well, VisiPics could certainly be faster and it hasn't been updated in a couple of years, but damn, it still does a pretty good job, I've been using it for years.

And yeah, even if you don't continue the project, the knowledge could be useful for something else in the future.
I guess the color there is mostly because of skin/penis color, though.

while on the software topic, is there a good program that will display selected images into a well ordered across all my monitors with minimum space waste?

>>4687013

Well, it doesn't sound too hard to make if nothing else. Care to elaborate?

>>4686736
I use awesome duplicate photo finder. it's pretty straight forward and works well. I've even had it pick up on artists redrawing old images. I do wish there was a way to flag images as false positives or something for future scans though

http://www.duplicate-finder.com/photo.html

>>4687013
Do you mean it should just tile them next to each other so as to minimize empty pixels?

I have an old ass folder full of hentai, with a lot of doubles, what software should I use to filter those doubles and clean them quickly?

that moment when hentai pics and data analytics merge, I´ll still fap to that hentai-graph tho!

>>4689092
I use Doublekiller and Visipics. Both are free and small downloads.

>>4689140
The thing about my folders is that there is a lot of raws,translated and licensed stuff that are the same, would doublekiller and visipics work with them?

>>4686713
You may want to look into some of the more classical solutions to this kind of problem.

# Removing exact duplicates #
Apply hashing function to each image and store each hash in a database next to a reference to each image. Then you can just search the database for files with duplicate hashes.

# Removing non exact duplicates #
For this you want an algorithm that tells you the similarity between two images. There are a number of these but structural similarity index is probably where yo want to start.

Additionally if the goal is to save hard drive space then you may also want to consider stripping image meta data and running them through a lossless image compressor.

>>4692232
>Additionally if the goal is to save hard drive space then you may also want to consider stripping image meta data and running them through a lossless image compressor.
This would probably be better achieved via an external script. Some formats, like png, have a massive spectrum of approaches to filter selection, bitdepth reduction, palette ordering, etc. The old pngout -r, then huffmixing can still beat many hours with something like pngwolf, then zopfli defluff and deflopt.

>>4686713
I want to fuck with this for the sake of fucking with it

link?

>>4686713
Wow this is more useful than the shit /g/ comes up with for the most part

Doublekiller uses matching hash numbers, or other selectable criteria. Visipics allows you to examine each file before deletion, it also uses a slider bar to specific how close multiple files match.

>>4686713
I wanna try it

>>4689148
Yep, if you relax the slider on visipics enough it'll give a ton of false "close-enoughs".