Section / Chapter headers are recognized based on font sizes only. There is no semantic recognition here!
Therefore (your example), if headers have no larger font size than body text, they cannot be detected.
The text content was not recognized either
As shown in the following figure
The "表1 基础指标" was not identified
This is a text, not a title!
pymupdf4llm == 0.0.27
file:
test.pdf
@JorjMcKie