docs: say background removal can drop objects a person holds (#6002)

The included people-only model can drop all or part of a held microphone, phone,
card or instrument. Measured against isnet-general-use and BiRefNet (general and
lite): those keep held objects far better but run 2.8x to 28x slower per frame,
so the default stays and the guide now names the case.

docs/guides/remove-background.mdx

@@ -97,6 +97,8 @@ subject, stable framing, and reasonable contrast from the background.
 Use another segmentation or masking tool when:
 
 - the subject is a product, animal, or other object;
+- the person holds something, especially a dark microphone, phone, card, or
+  instrument: the model can drop all or part of it;
 - fine hair crosses a busy background;
 - frame-to-frame edge stability must meet high-end VFX standards;
 - you need the original background reconstructed after removing the person.
@@ -126,6 +128,7 @@ cutout sits directly over its original source and edge color must match; use
 | Processing is slow                | Check `--info`; CPU processing works but is much slower than an available accelerated provider. |
 | The subject has a color halo      | Use `--quality best`, especially when layering over the original source.                        |
 | The “background plate” has a hole | Expected: it is inverse alpha, not an inpainted clean plate.                                    |
+| Part of a held object is missing  | The model matches people, not what they hold. Use another matting tool for that shot.           |
 
 The [CLI reference](/packages/cli#remove-background) lists every flag and
 supported input.