[Bee] Complex Excel layouts result in duplicate tables when converted to HTML

### Bug

This problem occured when converting a large customer Excel spreadsheet to HTML -- I'm not able to post it here because it contains customer information.

From an engineer that looked into the issue:
_The sheet is one large table with complex merged cells and docling seems to incorrectly fragment it into 293 separate tables (I think because of the merged cells). The section headers appear both inside the main 597-row table and again as separate 1-row tables which is causing the duplication, so it seems like a table-segmentation bug in docling's backend._

Can this problem be addressed?

### Steps to reproduce

...

### Docling version

docling: ~=2.96.1
docling-core: ~=2.78.0
docling-ibm-models: ==3.13.0
docling-parse: >=6.2.0, <7.0.

### Python version

Python: >=3.11, <3.13

Start by examining the table-segmentation backend and reproducing the Excel-to-HTML conversion with a workbook containing complex merged cells. Compare the generated tables with the source sheet; done means section headers are not duplicated and the large sheet is emitted as the correct table structure.

Written by the indexing model from the issue text.