The other night I took one more step toward becoming my father when I found myself watching the TV news before dinner. Thankfully it was PBS's NewsHour, so it wasn't quite so callow and shallow as some TV news. Like so much good content via public broadcasting, it was very in depth too. It was an interesting story about the future of facial recognition, and how computers have now come to identify facial features from even the most low-resolution and blurry images. It turns out those cheesy TV police dramas that magically improve the quality of a low-res video image with the click of a mouse... Well, that technology isn't actually that far off. What's interesting, though, is it's not that the pixels are improved so much as the computer is trained to learn what a real high-resolution face looks like at ultra-low resolution—so when it sees that low resolution image it can effectively match it to a higher resolution version. I think this is evidence that just when we think we've figured out the rules around digital imaging, someone much smarter than me comes along and changes the game. I wonder how this type of technology will impact the way we make pictures too? Will cameras eventually be smart enough to improve our digital images by drawing from ideal versions of similar images in its database? You never know. Check out the NewsHour report via the link below.