Pathovio
Back to the blog
EB-2 NIW

Scanning and PDF standards for immigration exhibits: resolution, OCR and file size

Set up a scanning and PDF workflow that produces legible, searchable, correctly sized exhibits for an EB-2 NIW filing before you assemble the packet.

9 min read TR
Share this article

Why scan quality matters to an officer, not just to you

An adjudicator reads a petition on a monitor, not on paper. Diplomas, award letters, and publication excerpts arrive as PDF pages that get zoomed in and out, panned across, and sometimes printed for internal review. A scan that looks fine on your own screen at a glance can become an unreadable block of gray noise the moment someone magnifies it to check a signature, a seal, or a date.

This is not only a matter of presentation. USCIS's electronic filing systems and RFE response portals enforce specific file size and format constraints, and documents that don't comply can be rejected outright or flagged for resubmission. A skewed page, a scan missing its OCR text layer, or a file that exceeds the size ceiling for a single upload all create friction at the exact moment you want your record to move smoothly.

More importantly, illegibility has a direct procedural consequence: if an officer cannot read a critical document — a foreign degree, a citation count, a signature block on a reference letter — the standard response is not to guess in your favor. It is to issue a Request for Evidence asking for a clearer copy, which adds months to the case and signals that the initial filing was not carefully assembled. Getting scan quality right the first time is a control you have; a follow-up request for legibility is time you do not get back.

Target resolution and color settings for each document type

Different exhibits need different scanner settings. Set these before scanning a batch, not after — rescanning a stack of reference letters because the DPI was too low wastes more time than checking the settings once.

| Document type | Resolution | Color mode | When to use color | |---|---|---| | Standard printed pages (reference letters, publications, transcripts, diplomas) | 300 DPI | Grayscale | Use color if the page has a raised seal, embossed stamp, colored letterhead, or a wet-ink signature that needs to be distinguishable from printed text | | Photographs or figures (award ceremony photos, lab images, charts with color coding) | 400–600 DPI | Color | Always — color loss can make a chart or photo useless as evidence | | Forms or documents with fine print (visa stamps, small tables, footnoted citations, dense certificates) | 400–600 DPI | Grayscale, or color if the form uses color-coded fields or a security seal | Use color when the issuing agency's seal, hologram, or stamp is part of what authenticates the document | | Signatures and notary stamps on any document | Same as the base document, minimum 300 DPI | Color if the stamp or ink is colored | Use color whenever a signature or stamp's authenticity may need to be visually assessed |

When in doubt, scan in color and grayscale versions and compare file size against legibility before deciding.

A quick test

Zoom the scanned page to 100% on screen. If a stamp, seal, or small-print citation is blurry or illegible at that zoom, increase DPI and rescan that page.

Producing a clean, straight, correctly cropped scan

Choose the right capture device

Use a flatbed or sheet-fed scanner rather than a phone camera whenever one is available. Scanners produce even lighting, consistent geometry, and predictable DPI; phone photos introduce perspective distortion, shadows, and inconsistent resolution that are hard to fix afterward. If a phone camera is the only option (for example, a bulky bound certificate), shoot straight-on in even light, disable flash, and use a document-scanning app that auto-crops and flattens perspective rather than a plain camera roll photo.

Check alignment and cropping

After scanning, inspect each page for skew — text that tilts even a few degrees will look careless and can affect OCR accuracy. Use the scanner software's auto-deskew function, or a PDF editor's straighten tool, and re-check afterward. Crop tightly to the page edges, but preserve a small margin so no text, seal, or signature line is cut off. Watch corners and margins in particular, since automatic crop tools sometimes clip page numbers or footers.

Verify contrast and surface quality

Zoom in on stamps, embossed seals, and signatures to confirm they remain visible against the background — faint elements can disappear entirely if the scanner's contrast or brightness settings are too aggressive. Rescan any page showing shadows along the binding edge, moiré patterns (common with glossy certificates or photographs), or margins cut off during feeding. A page worth rescanning is one you would not want an officer to squint at.

Making the PDF text-searchable with OCR

A scanned page is just an image until OCR (optical character recognition) adds an invisible text layer behind it. Without that layer, no one — officer or reader — can search, copy, or index the document electronically.

Running OCR

  1. Open the merged, cleaned-up PDF in an OCR-capable tool: Adobe Acrobat Pro (Scan & OCR / "Recognize Text"), ABBYY FineReader, or the free command-line tool OCRmyPDF.
  2. Set the language correctly if the source document is not in English.
  3. Run OCR on the whole file, not page by page — this keeps the output as a single searchable PDF rather than fragments.
  4. Save with a new pass only if you re-scan or replace a page; OCR layers do not update automatically when you edit a PDF.

Verifying accuracy

  1. Open the finished PDF in a standard viewer.
  2. Press Ctrl+F (or Cmd+F on Mac) and search for a distinctive phrase you know is on the page — a name, a date, an institution.
  3. Confirm the viewer jumps to and highlights that text.
  4. Spot-check two or three pages per exhibit, especially any with dense small print, stamps, or unusual fonts, where recognition errors are more common.

Searchable text lets an officer locate specific content quickly during review, and lets you cross-reference citations in your exhibit index or personal statement back to the exact page without relying on memory or manual page-flipping.

Combining pages into one PDF per exhibit and naming files

Each exhibit should exist as a single PDF, regardless of how many pages it contains. A five-page recommendation letter or a ten-page publication is one file, not five or ten. Scanning software and OCR tools generally output one file per scan batch — set the scanner or scanning app to append pages into one document, or merge them afterward using Acrobat's "Combine Files," a free tool like PDFsam, or the merge function built into most OCR utilities. Delete blank separator pages the scanner may insert between originals before merging.

Naming convention

Name each file to match the exhibit number and short description used in your exhibit index and table of contents — do not let the two diverge. A workable pattern:

Exhibit-[Number]_[Short-Description].pdf

Examples:

  • Exhibit-07_Letter-Dr-Smith.pdf
  • Exhibit-12_Publication-Nature-2021.pdf
  • Exhibit-19_Diploma-PhD.pdf

Use two-digit numbering (01, 02, ... not 1, 2) so files sort correctly in a folder listing once you pass exhibit nine. Avoid spaces if your e-filing system has ever mishandled them — hyphens or underscores are safer. Keep the description short enough that the file name doesn't truncate when displayed in USCIS's upload interface.

Cross-check before moving on

Open your exhibit index and, line by line, confirm that every listed exhibit number has a corresponding file with a matching name, and that no file exists without a corresponding index entry. A mismatch here is easy to introduce during merging and easy to miss until an officer notices it first.

Controlling file size without losing legibility

A high-resolution OCR'd scan can easily run several megabytes per page. Before filing, reduce file size deliberately rather than letting a tool guess.

Compress with the right tool

  1. In Adobe Acrobat, use File > Save As Other > Reduced Size PDF, or the PDF Optimizer for more control over image downsampling.
  2. If you don't have Acrobat, Ghostscript (free, command-line) can compress PDFs in batches — search for the current recommended command syntax, since options change between versions.
  3. Online compressors can work for non-sensitive documents, but avoid uploading anything containing personal identifiers, signatures, or seals to a service whose data handling you can't verify. Prefer local, offline tools for identity documents and government-issued records.

Downsample, don't destroy

Apply the resolution targets from the earlier table when downsampling — 300 DPI for text, up to 600 DPI for small print, seals, or photographs. Compression tools often let you set a maximum DPI ceiling; use that rather than an aggressive "high compression" preset, which can blur fine lines in seals, signatures, and small-font text without warning.

Recheck after compressing

After compression, reopen the file and zoom to 100%. Confirm that seals, watermarks, stamps, and any small print (dates, certificate numbers, footnotes) are still crisp enough to read without additional zooming. If any element blurs or pixelates, back off the compression level and re-export rather than accepting the smaller file.

Meeting USCIS's format and size limits for e-filing and RFE responses

USCIS periodically changes the per-file and total-package size limits, and the list of accepted file types, for both online filing accounts and mailed RFE responses. Do not build your exhibits around a number quoted in a guide, this one included. Before assembling your final packet, check the current limits on your USCIS online account's upload instructions, or on the filing tips page for the specific form you are using at uscis.gov. Note the limit at the time you file, since it can differ from what applied when you started preparing documents months earlier.

If a single exhibit exceeds the limit

When one exhibit — a long publication, a lengthy reference letter with attachments, a bulky patent file — exceeds the per-file limit after compression, split it into sequential parts rather than shrinking it until it becomes illegible.

  1. Divide at a logical break: end of a chapter, end of a letter, end of a distinct attachment.
  2. Name each part clearly and consistently, for example Exhibit-12_Part1of2.pdf and Exhibit-12_Part2of2.pdf.
  3. Update your exhibit index to show both parts under the same exhibit number, with a note indicating the split (e.g.,

Final quality check before filing

Before uploading exhibits or assembling a paper packet, run through every file one more time with fresh eyes — ideally after a break, since skimming a document you just built makes it easy to miss errors.

Checklist

  1. Page completeness and order — Open each exhibit PDF and page through it. Confirm nothing is missing, nothing is duplicated, and pages appear in the same order as the original document.
  2. No blank or stray pages — Delete blank scanner pages, accidental double-feeds, or test scans that got merged in by mistake.
  3. OCR search test — Press Ctrl+F (or Cmd+F) in the PDF and search for a distinctive phrase from the middle of the document. If it doesn't return a hit, the OCR layer failed and the file needs reprocessing.
  4. File names match the exhibit index exactly — Check spelling, exhibit numbers, and punctuation against your table of contents. A mismatch here creates confusion during review even when the content is correct.
  5. Clean opening in a standard viewer — Open each file in at least one viewer other than the one you used to produce it (e.g., a browser's built-in PDF viewer) to catch corruption, missing fonts, or rendering errors.
  6. Total packet size — Add up file sizes and confirm the total, along with each individual file, is within the current limits stated on the USCIS filing account instructions or the relevant form's filing tips page.

Treat this pass as the last checkpoint before submission, not a formality.