An algorithm that bothered for several days

Find an algorithm idea



For example, is probably a "against" word, composed of "1", and then there are some scattered around the "1", I now want to remove these scattered "1". My train of thought is to calculate the number of "1" connected together, and when its number does not exceed a constant, set them to 0 to fetch the saved chunks.

To put it simply, I just want to get how many "1" of a pattern composed of "1", which consists of how many "1" 。 How to write this algorithm, I algorithm is in is too rotten

Reply to discussion (solution)

Gaussian noise reduction algorithm

Take a look around each point to see if it's three steps in a row (you can have a 1 to take over). No, it's going to be deleted.

Take a look around each point to see if it's three steps in a row (you can have a 1 to take over). No, it's going to be deleted.

Started with recursive continuous check, the results found that the efficiency is too low ...

Gaussian noise reduction algorithm

Baidu, Google on the data is very little ah, can provide an example of the code link it?

I think can only use the graph correlation algorithm, the mathematics algorithm, cannot avoid the against small write I above the point manslaughter

I think can only use the graph correlation algorithm, the mathematics algorithm, cannot avoid the against small write I above the point manslaughter

Do you have any good recommendations?

Wood, not familiar with graphics
can go to graphic image/Machine vision Http://bbs.csdn.net/forums/ST_Image Ask

All I think about is using fonts to generate against diagrams, and then make a difference with the original, but it doesn't seem like a good solution.

Wood, not familiar with graphics
can go to graphic image/Machine vision Http://bbs.csdn.net/forums/ST_Image Ask

All I think about is using fonts to generate against diagrams, and then make a difference with the original, but it doesn't seem like a good solution.


Oh, oh, anyway,thx.

Your needs may be related to this http://bbs.csdn.net/topics/390438979

I've done this, and you're thinking exactly the same, and I'm like an array, and then I scan each character line by column, then count the percentages. In fact, the low efficiency of the scare

Your needs may be related to this http://bbs.csdn.net/topics/390438979

Worship!

Your needs may be related to this http://bbs.csdn.net/topics/390438979

Worship!

The principle is probably the same, the software said in the study, is nothing more than to increase the code library, I think the main difficulty is still in front of the removal of interference, extract a relatively pure block on the

