Facial Recognition & Counter-terrorism pre-9/11
In Friday's Washington Post there was a story about Facial Recognition technology being used at State Department. I was involved in early research there on this technlogy. Here is the link to the article:
Passport ID Technology Has High Error Rate (washingtonpost.com)
The rest of the story I sent to the editor:
Dear sir:
The story on facial recognition in today's Post was a delight to see. However, as is often the case in our government, those who deserve credit do not appear in the story. There was no mention of Dr. George Gunn, the scientist who built the lab for testing and established the usefulness of facial recognition at State Department's Bureau of Consular Affairs. Dr. Gunn's work was introduced to NIST's Biometrics Consortium, in the fall of 2000. A short time later, Mr. Frank Moss, who your article quotes, actually decided against accepting funds from the White House's Office of Counterterrorism Technology and collaborating with other government agencies to further biometrics work. I recall him saying, "Their money is more trouble than it is worth."
What was not more trouble than it was worth, at that time, was moving the facility where Dr. Gunn had tested on real VISA application images from Portsmouth, NH to Goldbug, Kentucky. This move, by most accounts, provided no benefit to the government or the taxpayer, but pandered to Kentucky politicians and some in the Republican controlled Congress who wished to punish a New Hampshire politician whose loyalty had been found lacking. Moreover, funding for the move, very likely came from VISA fees which Congress established after the first world trade center bombing for the purposes of shoring up our border security. No one said whether Mr. Moss, himself, was from Kentucky, but numerous natives of Kentucky had found their way into key positions under him, by that time.
From the results of the testing, it was determined that facial recognition technology had its shortcomings. It was recommended that people not build too much of a reliance upon it; but that it could, nevertheless, be used to significantly reduce costs within State's Diversity VISA program, where duplicate submissions for an immigrant VISA lottery had significantly increased workload and threatened the program's integrity. Standards for VISA and passport photographs eliminate some inaccuracies one might experience with common photographs, but even if only 60% accurate in detecting duplicates it could help detect duplicates and deter a significant workload (that had increased from 5-million applications to 11-million in a short period. Figuring a handling cost of $4 per application, cost savings would exceed $4 million per year and might save as much as $24 million per year if deterrence reduced applications to a number in line with more fraud-free earlier years. The business of sorting pictures and comparing pictures could also be greatly improved by this technology without the technology needing to be very accurate. Opportunity to pursue such savings did not create any interest.
One other thing: The reason this technology was brought into State Department was because a Massachusetts company, LAU Technologies, went to their Senator seeking guidance (it wasn't Ted Kennedy, so I think it was Sen. Kerry). A State Department Foreign Service Officer, who had been assigned to the Senator's office as part of some special program, sat in on the meeting with the senator and suggested the State Department. Things would not have gone further except for the persistence of a Ms Drury White in the Department's Small and Disadvantaged Business Procurement Office. She kept calling the Consular Bureau until Consular Bureau managers gave in and sent an underling to the meeting. An RFP was being written, by this same person, that would involve researching infrastructure technology issues. A subtask for facial recognition was added, and after that the system worked for a while.
I might also say that Dr. Gunn's lab would never have been built if he had been kicked off the task for disrespect when he browbeat the government task manager about not doing enough to advance the technology quickly.
Too bad that with American bureaucrats we have to have two World Trade Center Bombings for them to get on the stick. One would think that Pearl Harbor alone would have been enough.
Respectfully yours,
/S/ J Burford Fields
Passport ID Technology Has High Error Rate (washingtonpost.com)
The rest of the story I sent to the editor:
Dear sir:
The story on facial recognition in today's Post was a delight to see. However, as is often the case in our government, those who deserve credit do not appear in the story. There was no mention of Dr. George Gunn, the scientist who built the lab for testing and established the usefulness of facial recognition at State Department's Bureau of Consular Affairs. Dr. Gunn's work was introduced to NIST's Biometrics Consortium, in the fall of 2000. A short time later, Mr. Frank Moss, who your article quotes, actually decided against accepting funds from the White House's Office of Counterterrorism Technology and collaborating with other government agencies to further biometrics work. I recall him saying, "Their money is more trouble than it is worth."
What was not more trouble than it was worth, at that time, was moving the facility where Dr. Gunn had tested on real VISA application images from Portsmouth, NH to Goldbug, Kentucky. This move, by most accounts, provided no benefit to the government or the taxpayer, but pandered to Kentucky politicians and some in the Republican controlled Congress who wished to punish a New Hampshire politician whose loyalty had been found lacking. Moreover, funding for the move, very likely came from VISA fees which Congress established after the first world trade center bombing for the purposes of shoring up our border security. No one said whether Mr. Moss, himself, was from Kentucky, but numerous natives of Kentucky had found their way into key positions under him, by that time.
From the results of the testing, it was determined that facial recognition technology had its shortcomings. It was recommended that people not build too much of a reliance upon it; but that it could, nevertheless, be used to significantly reduce costs within State's Diversity VISA program, where duplicate submissions for an immigrant VISA lottery had significantly increased workload and threatened the program's integrity. Standards for VISA and passport photographs eliminate some inaccuracies one might experience with common photographs, but even if only 60% accurate in detecting duplicates it could help detect duplicates and deter a significant workload (that had increased from 5-million applications to 11-million in a short period. Figuring a handling cost of $4 per application, cost savings would exceed $4 million per year and might save as much as $24 million per year if deterrence reduced applications to a number in line with more fraud-free earlier years. The business of sorting pictures and comparing pictures could also be greatly improved by this technology without the technology needing to be very accurate. Opportunity to pursue such savings did not create any interest.
One other thing: The reason this technology was brought into State Department was because a Massachusetts company, LAU Technologies, went to their Senator seeking guidance (it wasn't Ted Kennedy, so I think it was Sen. Kerry). A State Department Foreign Service Officer, who had been assigned to the Senator's office as part of some special program, sat in on the meeting with the senator and suggested the State Department. Things would not have gone further except for the persistence of a Ms Drury White in the Department's Small and Disadvantaged Business Procurement Office. She kept calling the Consular Bureau until Consular Bureau managers gave in and sent an underling to the meeting. An RFP was being written, by this same person, that would involve researching infrastructure technology issues. A subtask for facial recognition was added, and after that the system worked for a while.
I might also say that Dr. Gunn's lab would never have been built if he had been kicked off the task for disrespect when he browbeat the government task manager about not doing enough to advance the technology quickly.
Too bad that with American bureaucrats we have to have two World Trade Center Bombings for them to get on the stick. One would think that Pearl Harbor alone would have been enough.
Respectfully yours,
/S/ J Burford Fields
posted by JB | 9:24 PM
0 Comments:
Post a Comment
<< Home