Image processing on bifurcation diagram to get small eps size -

I am creating a bizarre diagram (which is usually used in nonlinear mobility). Due to changes in stability in these diagrams, sudden change in topology is identified. These are sudden changes because one or more parameters pass through some important values ​​(s). Here's an example:

On the above figure, some image processing has been done to make the conspiracy more visually pleasant. A partition drawing typically has hundreds of thousands of points and the resulting eps file can be very large. Along with this, the alliance will move together simultaneously and there will be a bigger blot and the details will not be as per the figures given in the above link. Reducing the size of the marker in Matlab can be of little help, although the plots show more aesthetics.

Journal submission in latex format requires that the data be deposited in the EPS format. In my case, one of such data can result in about 6 MB in matlab and also in Gnuplot.

For the example above, 100,000 x values ​​are calculated for each R and one can think that the resulting EPS file will be very large, although the site shows some image processing, which Makes them more visually pleasing. Can anyone explain to me how to take action? I can not understand the explanation given in the "Summary" section.

Would the resulting image processing also reduce the size of the size?

Besides, any suggestions on reducing the file size of such a huge eps?

Thank you very much ...

You essentially using antialiasing You can. To do this, to calculate and histogram the values ​​(say 8) for something within the desired range and with high density of x values ​​(you should have 8 n bin for every x, some large enough For N) You can then get the average of every 8x8 (if you are using 8 values ​​of r for your column), you can get that column block for a single column of pixels that will look smooth with traditional conspice. Go to the next group of r values ​​and repeat.