Two versions of a scanned contract sit on the desk. One is the draft the client reviewed. The other is what they signed and returned. Something changed between them, a number, a clause, a date, but finding it by flipping back and forth, holding one page in memory while looking at the other, is a test of attention the human brain was not designed for. Differences hide in plain sight.
The eye skips over what it expects to see. A comparison tool shows what is actually different, not what memory thinks is there.
Comparing two scanned documents requires either OCR to make text selectable and comparable, or visual overlay comparison highlighting pixel-level differences between page images. WukongPDF's PDF Editor and PDF Compare workflows support both approaches, and the method below applies to any pair of scanned PDFs.
Making Scanned Documents Comparable With OCR
Scanned documents are images. To compare their text, the text must first be extracted. Run OCR on both PDFs to create searchable text layers. OCR output is not perfect, especially on low-quality scans, but it is sufficient for comparison because the tool flags differences and you visually verify each flagged difference against the originals.
Adobe Acrobat Pro's Compare Files tool works on scanned PDFs after OCR. It compares recognized text and highlights differences in the context of the page image. The comparison is as accurate as the OCR, catching most substantive changes while sometimes flagging recognition artifacts as differences. Manually review each flagged item to separate real changes from OCR noise.
Using Visual Overlay Comparison for Pixel-Level Accuracy
Visual overlay aligns the two scanned pages on top of each other and highlights differing pixels. This method is independent of OCR accuracy and catches every visual difference, changed numbers, moved commas, shifted margins, added signatures. The output is a difference map showing changed regions in a highlight color.
The limitation: visual comparison flags every pixel difference, including scan artifacts, slight positioning variations, and brightness differences between scans. These false positives must be reviewed alongside real differences. Visual comparison provides completeness, every real change is found, at the cost of precision, not every flagged difference is meaningful.
Using Desktop and Browser-Based Comparison Tools
Adobe Acrobat Pro's Compare Files tool is the desktop standard. It compares both text and visual appearance, highlights differences, and produces a side-by-side report. Upload both PDFs, run the comparison, and review results in the comparison view. The tool works on scanned documents after OCR.
Browser-based comparison tools offer similar functionality without installing software. Upload both PDFs and the tool aligns pages and highlights differences. Quality varies by tool, and scanned documents are the hardest case because the tool must handle image alignment and variation. Test a browser-based tool on a pair of known-similar scanned pages before relying on it for an important comparison.
| Comparison Method | Strengths | Weaknesses | Best For |
|---|---|---|---|
| OCR + text comparison | Identifies substantive text changes | Misses formatting changes, OCR errors flag false positives | Contracts, legal documents, reports |
| Visual overlay (pixel-level) | Catches every visual difference | Flags scan artifacts as differences, no semantic understanding | Signed documents, forms with handwritten changes |
| Manual side-by-side review | Human judgment, no false positives from tools | Slow, fatiguing, misses subtle differences | Short documents, final verification pass |
| Hybrid (OCR + visual) | Combines strengths of both methods | Requires two separate comparison passes | High-stakes documents where no change can be missed |
Handling Handwritten Changes and Annotations
Scanned contracts often contain handwritten marks, initials in margins, circled terms, strikethroughs with a pen, notes added after printing. OCR handles printed text but fails on handwriting, especially cursive. Visual overlay comparison catches these marks because the pixels differ between the two scans, but it cannot interpret what the marks mean.
Handwritten changes on scanned documents require human review, period. No automated tool can reliably interpret the legal significance of a handwritten strikethrough through a dollar amount on a scanned contract. The comparison tool flags the location of the change. A human determines whether the change is a negotiated revision, a stray pen mark, or a scanning artifact. The tool finds. The human judges.
Reviewing and Interpreting the Comparison Results
The comparison tool produces a report listing every difference. Work through the list from top to bottom. For each difference, open both PDFs to the relevant page and visually verify the flagged difference is real, not an artifact. Accept or reject each one. Accepted differences are the changes between versions.
Categorize accepted differences by type: text changes, number changes, formatting changes, additions, deletions. Number changes, dollar amounts, dates, percentages, are typically more significant than formatting adjustments. Review number changes first. They carry financial and legal consequences that formatting changes do not.
Creating a Change Summary for Stakeholders
Compile accepted differences into a change summary. List each change with page number, original text, and changed text. The summary gives stakeholders a clear picture of what changed between versions without requiring them to run the comparison or review the full difference report.
Keep the original comparison report as a backup. The summary communicates. The report documents. If a stakeholder questions a specific change, the report provides the page reference and visual evidence to answer the question definitively.