
	Managing images The images I am referring to are those normally captured when using document optical storage and retrieval application programs. For the purposes of this kind of programs the images are captured by means of an image scanner, and then stored in some optical medium. Usually an image is captured at 200 dpi, one bit per pixel. Then it is compressed according to recommendation T.6 of the ITU-T (ex CCITT) Committee. This recommendation is better known as group IV-2d compression scheme. This is a good compromise between the quality of the captured image and its disk space occupation, usually an average of 23 to 45 kB for the letter size document. The best choice to store an image is packing it into a TIFF-B file, the bilevel subclass of the TIFF format, as also recommended into RFC 1314. Please refer to the TIFF specification to know how a file has to be structured. I would point out here that the TIFF file format offers a tag to accomodate user information. This tag is the ImageDescription tag. In document optical storage and retrieval application programs this tag should be used to store the keys used in the database to refer to the image. This way each image is fully self documented, and can exist independently from the existence of the database. Furthermore a database can be built from scratch collecting the information provided into the ImageDescription tags. Such an image can be exported anywhere.

Managing images

The images I am referring to are those normally captured when using document optical storage and retrieval application programs.

For the purposes of this kind of programs the images are captured by means of an image scanner, and then stored in some optical medium.

Usually an image is captured at 200 dpi, one bit per pixel. Then it is compressed according to recommendation T.6 of the ITU-T (ex CCITT) Committee. This recommendation is better known as group IV-2d compression scheme. This is a good compromise between the quality of the captured image and its disk space occupation, usually an average of 23 to 45 kB for the letter size document.

The best choice to store an image is packing it into a TIFF-B file, the bilevel subclass of the TIFF format, as also recommended into RFC 1314. Please refer to the TIFF specification to know how a file has to be structured.

I would point out here that the TIFF file format offers a tag to accomodate user information. This tag is the ImageDescription tag. In document optical storage and retrieval application programs this tag should be used to store the keys used in the database to refer to the image. This way each image is fully self documented, and can exist independently from the existence of the database. Furthermore a database can be built from scratch collecting the information provided into the ImageDescription tags. Such an image can be exported anywhere.