Gemini Pro Vision: Document Digitization and OCR

Process this document image and extract all information with high accuracy:

**1. Full Text Extraction**:
- Extract ALL text from the document, preserving the original layout and hierarchy
- Maintain formatting: headers, sub-headers, bullet points, numbered lists
- Preserve table structures with proper row/column alignment
- Identify and extract text from stamps, watermarks, or annotations
- Handle handwritten text if present (note confidence level)

**2. Document Classification**:
- Document type (invoice, receipt, contract, letter, form, report, certificate)
- Language(s) detected
- Date on the document
- Issuing organization or individual

**3. Structured Data Extraction**:
- Key-value pairs (e.g., Invoice #: 12345, Date: 2026-01-15, Total: $1,234.56)
- Line items in tabular format
- Signatures and sign-off information
- Reference numbers, codes, or identifiers

**4. Output Formats** (provide all):
- Plain text (cleaned and formatted)
- JSON with extracted key-value pairs
- Markdown with preserved formatting

**5. Quality Assessment**:
- Overall OCR confidence score (percentage)
- List any areas where text was unclear or uncertain
- Suggest if higher resolution would improve accuracy

Process with maximum accuracy. Flag any uncertain extractions with [UNCERTAIN].