Dear Marina,
We agree in several points. As you observed, photons are not point particles, and they don't bounce, they pass through one another, without interacting, indeed. I am not aware about the experiment you mention, but I think that the three streams of photons don't "see" each other. The fact that photons pass through each other undisturbed doesn't contradict interference, in fact interference wouldn't be possible without this. The problem is that, if after interference, you detect the photons, they are found in positions distributed according to the Born probability obtained from the wavefunction which resulted from interference. It is as if the wave, initially widely spread, concentrates at the point where it is found. As I said earlier, when you measure position, you find the photon localized around a point, but this doesn't mean it is point particle, but just a localized wave. So, the wavefunction before observation is epistemic, it gives the probabilities, and after the measurement, it is ontological, it is what you find. There are more ways to consider the wave to always be ontological. The most direct way is to consider that, just before being measured, it changes its shape, to become an eigenstate of the measured observable. This picture is like this: (1) The wave is ontological. (2) It evolves unitarily between measurements. (3) At measurement it is projected, according to the projection postulate (and Born's probability). This is probably close to your idea of shape shifting. It explains some features, and if this answers all of your questions, it may be what you want. It is not very good when you try to think how this applies in spacetime, especially when entanglement is involved, but it challenges one's intuition even for only one particle (by particle I don't mean point particle). For some, this picture may be sufficient, for me is not. It may be just a matter of preference after all. I explained 5 years ago some of my reasons why I am not happy with it, and what I proposed in place, in this video, and this article. I hope this will help.
Best regards,
Cristi