Towards a Unified IU Environment: Coordination of Existing IU Tools with the IUE

Progress in the eld of Computer Vision/Image Understanding has long been hampered by the lack of standard software environments for research and application development. The Image Understanding Environment (IUE) is being implemented to provide a freely distributed standard software environment appropriate for a wide range of research and development activities. Nevertheless, we recognize that the IUE will not serve everyone's individual needs and that conversion to the IUE could be an expensive process for many sites. Therefore, we are investing signiicant eeort to make the IUE an open environment that can work in coordination with existing IU systems and tools. This paper introduces the work being performed at Amerinex A.I. to address the processing and data representation factors involved in this coordination. A long-standing problem in Image Understanding research has been the lack of established software standards or standard software environments in the IU community 4]. Historically, research groups have developed their own systems in response to speciic investigations , and these systems have then evolved within the institution to serve wider and wider processing requirements. Such an approach seriously impacts the eld of computer vision/image understanding in two areas: Research performed at one institution cannot easily be duplicated at external sites because of software and data structure incompatibilities. Tools providing standard implementations of established computer vision algorithms can not be shared between research sites or used to transfer research technology to application developers 1]. These two factors have a dramatic eeect on the ability of the vision community as a whole to maximize the eeectiveness of its research eeorts. The Image Understanding Environment (IUE) has been designed to alleviate these diiculties by providing an eecient, extensible class hierarchy designed for the needs of image understanding researchers. Nevertheless , we recognize that adoption of the IUE will not be universal, and that there is an enormous investment in existing IU software that must be supported as research environments evolve. Therefore, the IUE design committee and the IUE implementation team at Amerinex are devoting a great deal of energy to the model of a uniied IU environment. This will allow researchers to integrate existing systems with the IUE while providing a smooth migration path and maintaining maximum research productivity as the software environment evolves. There are four speciic requirements that need to be addressed in the IUE in the pursuit of this model of a uniied IU environment: Image data compatibility. …