This method relates to the compression of information contained in filled-in forms (O) by separate handling of the corresponding empty forms (CP) and of the information written into them (VP). Samples of the empty forms are pre-scanned, the data obtained digitized and stored in a computer memory to create a forms library. The original, filled-in form (O) to be compressed is then scanned, the data obtained digitized and the retrieved representation of the empty form (CP) is then subtracted, the difference being the digital representation of the filled-in information (VP), which may now be compressed by conventional methods or, preferably, by an adaptive compression scheme using at least two compression ratios depending on the relative content of black pixels in the data to be compressed.