University of Florida / Center for Computer Vision & Visualization

AIM Project - Adaptive Image Manager

Sponsor: DARPA/ITO Adaptive Computing Systems Program

Technology Transfer

1. Technology Transfer Opportunities

1. Due to its architecture independence, the AIM software could be configured to function with a variety of parallel and networked computers, thus realizing portability of AIM to various military, government, medical, and business applications, as discussed previously.

2. Lockheed-Martin's PAL processor, together with AIM, would permit high-bandwidth image and signal processing (2.5 GFLOPs per PAL-II board) using interfaces such as Khoros (currently supported). This would realize robust, near-optimal, high-bandwidth SIMD-parallel computing for a variety of military and medical applications, at a reason able cost (less than $4k per board projected for PAL-II).

3. Advanced reconfigurable computers based on field-programmable gate arrays (FPGAs) would be more feasible in practice, since AIM would provide a means for integrating such processors. Examples of such machines are Virtual Machine Works' emulator boards, and various reconfigurable signal processors developed by Lockheed-Martin Sanders using Xilinx XC3000 and XC4000 PCI-compatible FPGA development platforms.

Thus, results of the AIM study could immediately benefit military applications that required dynamic reconfigurability, especially as low-level tools for reconfiguring such processors are further developed and fielded.

Given UF's long history of success in product development and technology transfer (summa rized below), enhancement of product quality, scope, and delivery could feasibly be achieved in collaboration with the sponsor, subcontractors, and affiliated organizations. This would further support DoD's goals of increasing research and development capabilities of military, government, university, and industrial organizations.

Technology transfer to numerous areas is expected as a result of the AIM development effort. In addition to numerous military applications, domestic uses of the AIM algorithm development environment exist in areas of medical image processing, law enforcement (e.g., air port and crowd surveillance), drug interdiction (e.g., contraband detection), and environmental surveillance (e.g., pollutant detection).

UF is well versed in technology transfer, and has successfully participated in the practice of obtaining and licensing patents on algorithms, software, and hardware. For example, UF has pat ented or copyrighted numerous algorithms and software in the areas of automated target recogni tion, high-bandwidth sequential and parallel computing, and medical image processing. UF researchers have obtained and licensed numerous patents in areas of biochemistry and chemistry (e.g., Bioglass` and Gatorade`), physics and nuclear engineering (e.g., X-ray detection of bur ied land mines), food science, civil, electrical, mechanical, and coastal engineering, as well as copyrights on a wide variety of intellectual and artistic property. UF has realized over $9 million in royalties from patents and copyrights, with over $290 million grant and contract support for research and development in academic year 1994-95 (external research funding for UF's College of Engineering exceeded $40 million in 1995).

Substantial technology transfer to the defense and private sectors is expected from the current AIM effort, due to UF's affiliation with organizations such as DARPA, DoD-sponsored research agencies, various intelligence agencies such as CIA and NSA, as well as domestic agencies such as NASA, NSF, NIH, DOE, EPA, and numerous industrial firms. For example, the NASA South east Technology Applications Center (STAC) is located at and cooperatively managed by UF. Additionally, UF faculty and administrators staff and manage the Graduate Engineering Research Center at Eglin AFB, FL.

UF has a long and successful association with the Air Force Wright Laboratories (USAF/ WL), which would facilitate the rapid transfer of the AIM software to applications areas in the public and private sectors. UF is a member of the PAL consortium (including USAF/ WL and Lockheed-Martin, Inc.), which is developing the Parallel Algebraic Logic (PAL) series of processors. PAL-I, currently being marketed, produces over 385 MOPs (8-bit operations) per board, and PAL-II, currently being prototyped, is designed to achieve 2.5 GFLOPs (32-bit IEEE floating point operations) per board. Additionally, UF has agreements in place with USAF/WL for the processing of secure data at a Tempest facility located at Eglin AFB. Similar agreements are in place between UF and Lockheed-Martin, in Orlando, FL.

2. Ongoing Technology Transfer Paths

• Los Alamos National Laboratory (LANL) has been contacted regarding the AIM software. Correspondence with Jeff Bloch is in progress regarding use of AIM for controlling image and signal compression applications running on reconfigurable architectures.

  Additionally, Leon Alkalai's research group on fault tolerance, has expressed an interest in correspondence and receiving AIM updates. This would be congruent with Dr. Alkalai's ongoing research in fault tolerance for small spaceborne processors.

• Sandia National Laboratory, in particular, Dr. Hofstetler's research group, has expressed interest in sharing models of computation. They have detailed models of FPGA and SIMD processors, while we are developing medium-granular models that simulate various fault and failure modes.

• Harbor Branch Oceanographic Institution (HBOI) has expressed keen interest in applying AIM and reconfigurable processor technology to image compression applications. In particular, UF and HBOI are currently funded by the Office of Naval Research (ONR) to develop high-compression image transformations for increasing the exploitation of underwater acoustic communication channels for UAV telemetry. AIM technology could be used to extend apply this research to low-power on-board reconfigurable processors, versus power-consumptive DSPs.

This document is Copyright © 1997,1998 by UF/CCVV.
All rights reserved, except for use by the U.S. Government and its agencies or contractors.