Electron-muon mixing
In a muonic atom, the muon is like an electron, but more massive, which means it 'orbits' at 1/200 the radius an electron would. In theory you could put together mini molecules this way and have a tiny, tiny person or other object. Small problem: muons decay after just 1.5 microseconds, so you have to be very fast with your chemistry set.
Now let's suppose a moment that someone makes an Honest-To-God physics breakthrough, and can manipulate the fundamental forces that determine how heavy a lepton is. Just as the electromagnetic and weak magnetic forces were thought to be different until a way was uncovered to consider them as one, perhaps you can figure out a way that you can "tune" a lepton to any point between being an electron or a muon. You are able in this way to broadcast some hitherto-unknown particle (the 'shrinkon') into the region of space where a person is, and their electrons nearby become a little muon-ish and so they get smaller, but not so muon-ish as to begin decaying rapidly.
Now in all odds the only way to do this even if it can be done is to vaporize the person into a chamber hotter than the core of the hottest star (since this doesn't seem to have made itself known in stellar evolution), but the point is, we don't know that has to be true. Maybe there's a simple trick, just waiting to be worked out!
The shrinkons have to be something that have escaped the attention of every particle physicist - how I don't know, but I assume they're really heavy. Probably they should decay instantly, but let's suppose they take a while, and you can just blast them into a person with a particle accelerator, and then they emit some kind of gauge field that alters leptonic number around them in the body. Do it fairly uniformly and you have a mini-person, with the same nuclear weight, and we are not going so far into muonic character that the electrons have a significant weight, so to three or four places their mass is unchanged.
Note that the mini person will shrink things on contact, because they are emitting this hypothetical field. However, if we suppose it takes many, many particles embedded in the person, the field may project only a fraction of a millimeter. It can't be much less than that, because the epithelium of the skin is always regenerating from scratch, and without particles embedded in it your subject would quickly get full thickness skin, which would look odd, and the topology issues confuse me. But if they drink a glass of water, the water molecules approaching their intestinal lining shrink to match the rest of their body, and are absorbed like they would be in a full size person. Eating works the same way. Nonetheless, you have the formidable issue that the food contents of their intestine receive much less shrinkage if the range is short. But the more they are shrunk, the less the range needs to be ... I think this is a practical problem but perhaps manageable. They certainly feel like they have to eat and excrete a large volume (relative to their shrunken body); their intestines can do the biochemical work quickly but the physical bulk of the material is an issue. You can tune that down if you accept that they shrink your hand while they grab hold of it, and then it expands to full size. I feel like the geometry should get "altered" during this process depending on exactly how they approach, so that ought to be painful.
I have no idea at all about whether semi-muonic matter would absorb light at a wavelength proportional to its size, or whether their increase in mass compensates for that, or whether the light entering would be shrunk like eaten food. You could say they are blind to all but UV light if you want, but there's about as much case to be made for them seeing normally.