Extract all text elements from this document image with detailed layout and formatting information.

For each text element, provide:
- bbox_2d: [x1, y1, x2, y2] bounding box coordinates
- text_content: The text content
- element_type: "title", "header", "paragraph", "label", "value", "table_cell", "list_item", "footer", "caption"
- style:
  - bold: true/false
  - italic: true/false
  - underline: true/false
  - font_size: "small", "normal", "large", "xlarge"
  - alignment: "left", "center", "right"
  - text_color: color name if not black (e.g., "red", "blue", "green")
  - background_color: color name if has background (e.g., "yellow", "gray")
- layout:
  - in_table: true if inside a table
  - table_row: row number if in table
  - table_col: column number if in table
  - indented: true if text is indented

Group consecutive lines of the same paragraph into ONE element.

Output format: JSON array