ImSim: Image Similarity

[n00m]

Just admit that your algorithm doesn't work that well already
Or give a solid formal definition of "similarity" and prove that your
algo works with that definition.

Remind me this piece of humor:

One man entered a lift cabin at the 1st floor,
lift goes to the3rd floor, opens and ... it's empty!
Physicist, Chemist and Mathematician were asked:
what happened to the man?

Physicist: he was squashed to the floor by acceleration!

Chemist: he was vaporized by some acid gases!

Mathematician: hmm... Let's call a lift *empty* if
there is inside of it no more than *1* man.

 

[n00m]

So, my current very strict definition of similarity is:

-----------------------------------------------------------
2 pics are similar if my script gives for them value < 20%,
otherwise the pics are not similar.
-----------------------------------------------------------

It is left to study possible transitivity of similarity.


==========
LOL

 

[Mel]

But funny thing takes place.
At first thought it's a false-positive: some modern South East
Asian town and a lake somewhere in Russia, more than 100 years
ago. Nothing similar in them?

On both pics we see:
-- a lot of water on foreground;
-- a lot of blue sky at sunny mid-day;
-- a bit of light white clouds in the sky;

In short,
the notion of similarity can be speculated about just endlessly.

Exactly. That's the kind of similarity I would call valid. That's what my
algorithms, if I ever finished writing any, would be looking for.

Mel.

 

[Ian Kelly]

Remind me this piece of humor:

One man entered a lift cabin at the 1st floor,
lift goes to the3rd floor, opens and ... it's empty!
Physicist, Chemist and Mathematician were asked:
what happened to the man?

Physicist: he was squashed to the floor by acceleration!

Chemist: he was vaporized by some acid gases!

Mathematician: hmm... Let's call a lift *empty* if
there is inside of it no more than *1* man.

That reminds me of another joke I once heard.

A logician, a biologist, and a physicist are watching an empty house.
They observe two people going into the house, and then some time later
three people emerge.

The logician observes, "Our premise must have been incorrect. The
house was not initially empty."

The biologist speculates, "Perhaps the two people whom we saw go into
the house have reproduced while inside, resulting in the third."

Finally, the mathematician concludes, "I don't know about that, but I
do know that if one more person goes into that house, it will be
empty!"

This has nothing to do with Python, though.

 

[n00m]

@all and just in case.
Also see my TiRG project (since 2011-01-31):
http://sourceforge.net/projects/tirg/
It's for detecting and localizing textareas in raster graphics.
Among its files there is a python script -- absolutely working.
Enjoy to do with it whatever you like -- it's my public domain.

And again and this theme is a rich field for speculating about,
over "What on the earth is the text?" question.
Is 50 in a row letters 'o' some text or is it just a part of
some fancy Indian ornament? And so on, etc.

PS
Don't confuse it with OCR. It's 2 different beasts.
===