About

Barcode misreads cost the retail and logistics industries billions annually in inventory discrepancies, shipping errors, and compliance failures. This tool extracts encoded data from barcode images by converting pixel patterns into bar-space width sequences, then matching those sequences against standardized encoding tables for formats including EAN-13, UPC-A, Code 128, Code 39, ITF, and Codabar. It applies adaptive thresholding and multi-angle scanline analysis to handle real-world image conditions: uneven lighting, rotation up to 15°, and low contrast. The extraction is performed entirely client-side. No image data leaves your browser.

Limitations: 2D barcodes (QR, Data Matrix) require separate tooling. Severely blurred or occluded barcodes below approximately 100px width may not resolve. The decoder assumes standard module widths per ISO/IEC 15420 (EAN/UPC) and ISO/IEC 15417 (Code 128). For thermal-printed labels with ink spread, results should be cross-verified against the check digit, which this tool validates automatically.

Formulas

The decoder converts pixel luminance into binary bar-space patterns. Grayscale intensity for each pixel is computed as a weighted sum per ITU-R BT.601:

L = 0.299 ⋅ R + 0.587 ⋅ G + 0.114 ⋅ B

where R, G, B are the red, green, and blue channel values (0 - 255). Adaptive thresholding uses a local window mean to binarize the scanline.

For EAN-13 and UPC-A, the check digit C is validated using the modulo-10 algorithm:

C = (10 − (n−1∑i=0 d_i ⋅ w_i) mod 10) mod 10

where d_i is the i-th digit, and w_i alternates between 1 and 3 for EAN-13 (positions indexed from the rightmost digit).

For Code 128, the checksum uses modulo 103: the start code value plus each data character value multiplied by its 1-based position index, all reduced mod 103.

Bar width decoding normalizes measured pixel widths against the known module count per character. For EAN-13, each digit is encoded in 7 modules comprising 2 bars and 2 spaces, with width ratios compared against the L-code, G-code, and R-code lookup tables.

Reference Data

Format	Standard	Characters	Length	Check Digit	Common Use
EAN-13	ISO/IEC 15420	Digits 0-9	13	Modulo 10	Global retail products
EAN-8	ISO/IEC 15420	Digits 0-9	8	Modulo 10	Small retail items
UPC-A	ISO/IEC 15420	Digits 0-9	12	Modulo 10	North American retail
UPC-E	ISO/IEC 15420	Digits 0-9	8	Modulo 10	Compressed UPC for small packages
Code 128	ISO/IEC 15417	Full ASCII (128 chars)	Variable	Modulo 103	Shipping, logistics, GS1-128
Code 39	ISO/IEC 16388	A - Z, 0-9, 7 specials	Variable	Optional Mod 43	Automotive, defense (MIL-STD-1189)
ITF (Interleaved 2-of-5)	ISO/IEC 16390	Digits 0-9 (pairs)	Even count	Optional Mod 10	Carton labels, distribution
Codabar	ANSI/AIM BC3	0-9, 6 specials	Variable	Optional	Libraries, blood banks, FedEx airbills
Code 93	AIM BC5	Full ASCII	Variable	2 check chars (Mod 47)	Canadian postal, logistics
MSI Plessey	Proprietary	Digits 0-9	Variable	Mod 10 (Luhn) or Mod 11	Warehouse shelf labeling
Pharmacode	Laetus	Numeric (3-131070)	Variable	None	Pharmaceutical packaging
GS1-128	ISO/IEC 15417 + GS1	Full ASCII + AIs	Variable	Mod 103 + AI validation	Supply chain, SSCC, expiry dates
ISBN-13	ISO 2108 (EAN-13 subset)	Digits, prefix 978/979	13	Modulo 10	Book identification worldwide
ISSN	ISO 3297 (EAN-13 subset)	Digits, prefix 977	13	Modulo 10	Serial publications

Frequently Asked Questions

The tool supports all major 1D barcode symbologies: EAN-13, EAN-8, UPC-A, UPC-E, Code 128 (including GS1-128), Code 39, ITF (Interleaved 2-of-5), and Codabar. These cover over 95% of barcodes encountered in retail, logistics, healthcare, and library systems. The native BarcodeDetector API (available in Chromium-based browsers) may additionally detect Code 93 and other formats.

The decoder requires a minimum barcode width of approximately 100 pixels for reliable extraction. Critical factors include contrast ratio between bars and spaces (minimum 40% difference recommended per ISO/IEC 15416 print quality standard), image sharpness (Gaussian blur with sigma above 2 pixels will degrade results), and even lighting across the barcode region. The tool applies adaptive thresholding to compensate for uneven illumination gradients up to approximately 30% variation across the symbol width.

Yes. The extraction engine scans at multiple angles: 0°, ±5°, ±10°, ±15°, and 90° (for vertically oriented barcodes). This covers the vast majority of real-world label photos. Barcodes rotated beyond ±15° from horizontal or vertical may require the image to be pre-rotated before uploading. Perspective distortion (keystoning) from angled photography is not corrected; photograph barcodes as perpendicular to the camera plane as possible.

No. All processing occurs entirely within your browser using Canvas pixel manipulation and JavaScript. The image never leaves your device. No network requests are made during extraction. This makes the tool suitable for sensitive inventory labels, medical specimen barcodes, or proprietary product codes where data confidentiality matters.

A check digit mismatch means the extracted digit sequence fails its format's validation algorithm (e.g., modulo 10 for EAN/UPC). This typically indicates partial barcode damage, ink spread on thermal labels altering bar widths, or low image resolution causing adjacent bars to merge. The tool still reports the raw extracted value so you can identify which digit is suspect. For EAN-13, the check digit is position 13 - compare it against manual calculation using the alternating weight-1/weight-3 sum method.

UPC-A is a 12-digit subset of EAN-13. Every UPC-A code can be represented as EAN-13 by prepending a leading zero. The encoding structure is identical: 7 modules per digit, same guard patterns, same modulo-10 check digit algorithm. If the tool reports a 13-digit EAN-13 starting with 0, the underlying code is effectively a UPC-A barcode. Country prefixes 000-019 and 060-139 are assigned to the United States and Canada.