jerryfreak, there's a problem with this in terms of basic physics. The range of audible frequencies has a 1000:1 range of wavelengths because of the familiar 20 - 20,000 Hz business. But to block a sound wave, an object needs to be at least half the wavelength of that wave. So unfortunately, any baffle, etc., that you might practically use could only be effective at high frequencies.
If you used a baffle six inches in diameter, for example, the waves for any sound beneath about 1100 Hz would simply flow around it, and so forth. That might sharpen the localization in your stereo recordings and it's a worthwhile experiment for its own sake, but it won't turn your omnis into some other directional pattern.
There is currently a product being marketed by a Chinese microphone manufacturer which claims to absorb the sound waves from behind a directional microphone. But the product is just a curved piece of metal a few inches across, lined with sound-absorbing foam. That might have an effect above, say, 2 kHz, but certainly not below.
The way most cardioid microphones work nowadays is that they have a front and a rear sound inlet, with an acoustical delay line in the rear inlet, arranged so that sound arriving from the front is reinforced while sound arriving from the rear is canceled. But it's actually more complicated than that, because the entire internal architecture of the capsule (the internal volume of air) has to optimized for that purpose, and the membrane tension has to be set right for it (lower than with an omni), or else the frequency response will be all wrong.
In other words there's no real hope of transforming an omni into a usable directional microphone, unless the capsule was designed in advance to make this transformation itself, i.e. unless it was designed and built to be a two-pattern switchable microphone.