Mixed Image Format Environments: Comparisons, Tradeoffs, and Issues Booz*Allen&Hamilton Inc. Abstract As the extent of the networked image environments grows, it becomes increasing likely that multiple image formats will be present within a network. This paper addresses the general evolution of this environment through a spec$c examination of three image formats pew e-3 L7 nronomt. ninitn7 rmncro PAmlrr.rrr;nntinr. ;n y I” “C y, COCI‘G. Y‘&“U‘ ur‘u&r ““,,‘,,‘yI,“C~C‘“,‘ 6,‘ Medicine (DICOM), Image Processing and Interchange (IPr) -Image Interchange Format (IIF), and the National Image Transmission Format (NITF). For each format, its approach to image format development and data encapsulation are both presented and compared to the other formats, and distinguishing capabilities and limitations are identiJied. A possible integration scheme for IIF and DICOM is presented and the relationship of NITF to IPI-IIF and DICOM is ident$ed. Since a mixed image format environment could include acquired imagery, reference imagery material, or even document imagery, generalized interoperability will be a requirement offuture systems. Results are generalized to provide an indication as to how this problem may be addressed. 1: Introduction The mixed image format environment continues to confront and confound the architects and implementors of all scales of imagery information systems or networks today. This paper presents a number of issues associated --.ZL,- ..-:.-- .~-..,L:.-l- 1.----. d? -.....- I- 1.. - -:..-a. WILL usmg mulriple image rormats m a smgie information system. Three formats are specifically examined because of their likelihood to be included in imagery systems: Digital Image Communication in Medicine (DICOM), Image Processing and Interchange (IPI) - Image Interchange Format (IIF), and the National Image Transmission Format (NITF). DICOM is the American College of Radiology (ACR)/National Electrical Manufacturers Association (NEMA) standard for the archiving and communication of medical imagery. IPI-IIF is the International Organization for Standardization (ISO) standard for archiving and processing of multidimensional imagery. NITF (MIL- STD-2500) is the United States military standard for the transmission of defense-related imagery. Section 2 presents a case for the use of multiple image formats within information systems architectures. A summary of three example formats is given in Section 3 and Section 4 compares the three standards to yield an understanding of what each would contribute to a mixed format environment. On the basis of this analysis, an integration scheme is presented in Section 5 for the simultaneous use of all three standards in a common infnm&nn ~xrn+~m G&inn /; ,W.P.SP~+C +h.a crc.n.cwol IIIL”IIIIU.I”II “J “IVIII. Y”YCl”ll ” p..o”AIw u1u ~U”V’U characteristics of image formats and relates this to the United States Imagery System (USIS)l use of image formats. A summary is presented in Section 7. 2: Multiple Image Format Usage When multiple file formats are used in a common environment, obvious interoperability problems exist. A common reaction to this problem is to seek the single file format that will meet all needs-the ideal file format. This format would be capable of collecting, storing, searching, and transmitting data between systems or around the world. It would meet the domain-specific needs of the medical, military, commercial, and casual user. Finally, it would be equally applicable as either a standalone file or as content within a content architecture file format-such as the Office Document Architecture and Interchange Format (ODA). However, within the domain with which most are most familiar, it is typical to create subsets of our favorite standard (called profiles) that are tailored to our own needs. In the case of NITF, these are referred to as ~~ f- conrormance ieveis. in the case of DKXM, they are modalities. While some data may be moved among these profiles, they are not generally interoperable. A strict application of the ideal format argument would lead one to want a single profile for everything. While this is obviously extreme, it leads one to question the advantages of a “single” file format, since the use of a single format would still allow a lack of interoperability. Applications today support at least three format types for their image data: storage/transmission, display, and printer. This is true even though only the storage /transmission is considered an image fotiat. Over the past few years there has been a convergence of printer ’ The United States Imagery System (LJSIS) is a global imagery information system developed by the U.S. Department of Defense’s Central Imagery Office (CIO). O-8186-7560-8/96 $5.00 0 1996 IEEE Proceedings of IMAC ‘95 269 Proceedings of the 4th International Conference on Image Management and Communication (IMAC '95) 0-8186-7560-8/95 $10.00 © 1995 IEEEformats to the point where Postscript is the dominant format for document printing. In the areas of display, a few formats dominate the market, typically driven by CPU hardware implementation. This trend will most likely continue as the manufacturers of displays and printers want the widest possible markets for their .---L.-r- T-L.‘- _--_ _ -_-__^^ L-,. --cl-,.-.. CL- ^^^^ c--CL,. pIuUUcLY. llllS GuII”cIg;F;IIw llilb uut “GGII UK ulsci I”1 LUG area of storage/transmission formats. In fact, format creation continues to flourish as new analytical and data handling requirements are discovered and captured in a format. Applications then bear the burden of translation when it is required. In some cases, this is handled through the augmentation of existing formats while in other cases completely new formats are created. It must be acknowledged, however, that the display, printer, and storage/transmission tile format problems are fundamentally different. A display format reflects the underlying hardware for the efficient display and real time manipulation of the pixel data. A printer format