What is DJVU? DjVu Document Image Format Explained

What is DJVU?

DjVu (pronounced deja vu) is a compressed image format designed primarily for storing scanned documents, especially those combining text, line drawings and photographs. It was developed by researchers at AT&T Labs in the mid-1990s and later commercialized by LizardTech.

DjVu separates a scanned page into layers, typically a high-resolution text/line mask and lower-resolution background and foreground images, then compresses each appropriately. This layered approach produces files that are often much smaller than equivalent PDFs or TIFFs while keeping text sharp.

How DjVu Works

DjVu, pronounced "deja vu," is a format built for one specific job: storing scanned documents at very small file sizes. Its key insight is that a scanned page is not one uniform image but a mix of very different content, sharp black text and line art on top of smoother, colored background and photographs. DjVu treats these differently instead of compressing the whole page with a single method, which is why it can be dramatically smaller than an ordinary image of the same scan.

It does this with layered separation. A DjVu encoder splits a scanned page into a foreground layer (typically the crisp text and drawings, which need high resolution but few colors) and a background layer (photographs and paper texture, which can tolerate lower resolution). Each layer is then compressed with a method suited to it, and a mask determines which pixels come from which layer, so text stays razor-sharp while the background is stored cheaply.

For the text layer, DjVu uses pattern-matching compression: it recognizes that a scanned page repeats the same letter shapes over and over, so it stores each distinct shape once and then records where copies appear, much like a dictionary of glyphs. This is extremely efficient for pages of printed text. The format also uses wavelet-based compression for the continuous-tone background, and it supports progressive rendering so a page can begin displaying before it has fully downloaded.

Beyond the image layers, a DjVu file can carry a hidden text layer produced by OCR, so the document becomes searchable and its text selectable even though what you see is a picture of the original page. It also supports multi-page documents, embedded thumbnails, and an outline, making a single DjVu file behave like a complete scanned book.

History and Standardization

DjVu was developed in the late 1990s at AT&T Labs by a research team including Yann LeCun, Leon Bottou, and Patrick Haffner. Their motivation was to make it practical to put scanned books, historical archives, and image-heavy documents online at a time when bandwidth and storage were scarce and a color scan of a single page could be enormous. DjVu's layered approach could shrink such scans far below what general image formats managed.

The technology was later handed off to LizardTech and subsequently other stewards, and an open-source implementation, DjVuLibre, kept the format freely available and readable across platforms. This open tooling is what allowed digital libraries and archives to adopt DjVu without being locked to a single vendor.

DjVu found its strongest home in digital libraries and scanned-book archives, where its compression advantage for text-heavy scans was decisive. It never became a mainstream consumer format, however, and as PDF improved and storage grew cheap, DjVu settled into a specialist niche. It remains valued in archival and academic contexts but requires a dedicated viewer or plugin that most people do not have installed by default.

DjVu Compared to PDF

The natural comparison is with PDF, and for scanned documents specifically DjVu often produces smaller files thanks to its foreground and background separation and glyph-based text compression. For a large archive of scanned pages, that size difference adds up, which is exactly the case DjVu was engineered for. When every scanned page matters and storage or bandwidth is tight, DjVu can still hold a real edge.

PDF wins decisively on ubiquity and versatility. Essentially every device can open a PDF without extra software, while DjVu almost always requires a dedicated viewer or browser plugin. PDF is also a general document format that handles born-digital text, vector graphics, forms, and interactive features, whereas DjVu is specialized for raster scans. For sharing a document with an arbitrary recipient, PDF is the safe default.

In practice the choice comes down to context. DjVu makes the most sense inside a controlled archival workflow where the size savings on large scanned collections justify installing the right viewer, and where searchable OCR over scanned images is valuable. For everyday distribution, or any audience that cannot be assumed to have a DjVu reader, converting to PDF is usually the pragmatic move despite the larger file.

MKV Technical Specifications

DJVU vs Other Document Formats

DjVu compresses scanned pages efficiently but has far narrower software support than PDF or TIFF.

Advantages & Disadvantages

Common Use Cases

DjVu is used where scanned documents must stay small yet readable.