The Rescue(63)

“The works?” said Reeves.

“You’re the only person that has expressed an interest in my special project.”

“Am I that transparent?”

“Forward-thinking,” said Watts. “Everyone else is satisfied with the status quo. Who’s the sidekick?”

“The sidekick is Matt Kincaid, my deputy,” said Reeves. “He doesn’t know about your improved-range facial-recognition upgrade, and we would like to know a little more about it.”

“Long or short version?”

“Short, please,” said Reeves. “We have our hands full today.”

“I’ve been practicing my elevator pitch.”

Reeves turned to Kincaid. “The elevator used to stop at every floor.”

“Funny secret agent man,” said Watts. “Here we go. Current 3-D FR software measures the distances between eighty universally recognized nodal points on the human face, or as many as it can detect, comparing the captured face print to the face print we feed into the system. It’s over ninety-five percent accurate within a fifteen-yard range, which isn’t great, and its accuracy degrades exponentially beyond that. It’s an image-resolution issue.”

“I’ve read about using higher resolution or pan-tilt-zoom cameras to solve that,” said Kincaid.

“Deputy Dog did his homework. Very nice,” said Watts. “Coaxial-concentric PTZ cameras solve the problem, expanding that range to fifty yards, but now you’re talking about a complete overhaul of a citywide camera system the city didn’t want to pay for in the first place.”

“So you tweaked the software to improve the range?”

“Tweaked is kind of a crude word for essentially rewriting the program.”

“Sorry,” said Kincaid. “Your software enhances the image resolution?”

“I considered a variation of image resolution but scrapped the idea when something far simpler came to mind,” said Watts. “My modification maps fewer nodal points per digital frame but randomly selects different nodal points for each frame analyzed. Current technology defaults to the same nodal points for every frame, which is great for static, close-up targets. My change should allow the current camera system to process moving targets at thirty yards, with a sixty to seventy percent accuracy.”

“That’s impressive,” said Kincaid.

“I’d like to get that number to ninety percent. Then we’d have something special.”

“Then you’d quit and go to work for yourself,” said Reeves.

“It’s all proprietary work on behalf of the JRIC,” said Watts. “Though I’m sure I could recreate it if I had to.”

“Are you ready to put it to the test on the streets?” said Reeves.

“I was hoping you’d ask,” said Watts. “When your facial-recognition requests hit my desk this morning, I thought it would be a perfect match. And the timing is right. I don’t have any major surveillance initiatives running at the moment. If anything higher priority comes in, which is pretty much anything other than what you submitted, I have to switch back to the current software package.”

“I understand,” said Reeves. “We’re looking at forty-eight hours, maybe less, to pursue this.”

“It’s an interesting set of targets,” said Watts. “Do you think they’ve mapped the camera zones?”

“That’s why I’m here. They’ve proven to be a countersurveillance-savvy group.”

“In other words, they ditched you guys.”

“Twice,” said Kincaid.

“Damn,” said Watts.

“It wasn’t the proudest moment of my career,” said Reeves. “Which is why I really want to find them.”

“You came to the right place. Even if they’re using some kind of camera-zone mapping app, my modification will trip them up,” said Watts. “The apps I’ve tested map safe passage around the detection zones by overestimating detection ranges. Double, sometimes triple the known effective ranges of our software.”

“But not your modification,” said Reeves.

“They’ll pass right through the new detection ranges without suspecting it.”

“What’s the downside?” said Kincaid.

“Based on the accuracy percentages, you’re going to see some false positives.”

“I can live with a wild-goose chase or two,” said Reeves. “Let’s do it.”

“Excelente, amigos,” said Watts, sliding a piece of paper and a pen across the desk. “I’ll need your signature authorizing the experimental use of the software in conjunction with your requests.”

Reeves took the pen and quickly read the document, which had already been signed by Angela.

“Looks like you’ve been expecting me all morning.”

“Angela has screened every facial-recognition request that has come into the JRIC over the past three weeks,” said Watts. “She’s really excited about the software.”

“No wonder she was happy to see me,” said Reeves, signing the document.

Watts stood up and shook their hands.

“What’s your guess on how long it’ll take?” said Reeves.

“If they’re sneaking around the city right now, we’ll start getting hits within the hour. The software modification doesn’t leave much safe space in LA. If they’re playing it smart and staying off the streets like I would? You know the deal.”