Tag Archive: science
We recently wrote about unblurring blurry photos a couple weeks ago. Now comes video (embedded above) showing real-time, dynamic, and easy-to-do insertion of fake objects into any photo. This is part of research led by Kevin Karsch, a PhD student at the University of Illinois / Urbana-Champaign. The user makes a bare-bones sketch of objects and lighting in the photo, and then drags an object into the image. The software then realistically places the object into the photo with proper lighting and collision. It’s difficult to describe, so you really should watch the video. Look for animations going through photos, bouncing off walls, casting proper shadows, and interrupting complex light patterns in a very natural way. Ordinarily, this sort of composite work would take expertise and hours, but the narrator says that one example in the video was done by a novice user in just 10 minutes.
I found this video via James Fallows’ blog at the Atlantic. Fallows calls this the latest, “What hath God wrought?” moment for technology and links to a short essay worrying about the democratization of image manipulation tools.
I have incredibly fond memories of watching James Burke‘s late-1970s BBC series Connections late at night on PBS during middle school. It’s a strange show, a meandering narrated documentary series about the history of science that draws well-known and obscure connections between major events in the past and present. The show is borne from a connected worldview of technological progress, “that one cannot consider the development of any particular piece of the modern world in isolation. Rather, the entire gestalt of the modern world is the result of a web of interconnected events, each one consisting of a person or group acting for reasons of their own motivations (e.g. profit, curiosity, religious) with no concept of the final, modern result of what either their or their contemporaries’ actions finally led to.”
Now, the whole series is available to watch on youtube. Give each episode some time…they’re slow at first, but by the end of each episode, you’ll never have expected to have arrived where you ended from where you started.
Eye tracking has emerged as an important part of measuring audience engagement and user experience on the internet. Devices record where a user’s eyes fall when looking at a website and the data from those experiments guide the future design of the website. Via bloggasm, I came across a look into how users view photos as web content. And while the data is not directly applicable to news websites or journalism, it does confirm what photographers already know: relevant photos of people increase user involvement with content. Interestingly, the findings also show that stock photos of generic people and situations get almost no attention from the user.
We’re a ways off from hand-held cameras that can do it, but the future of photography will involve pictures in which the depth-of-field and focus and camera position each can be adjusted reliably and with quality in post-production. It’s a complex mathematical and computational problem, but the power is within reach.
So you want to influence the future of photography? Well, you gotta build a camera, ’cause this future isn’t for sale, yet.” -FuturePicture.com
Two enterprising photography enthusiasts have taken a page from MIT‘s and Columbia‘s and Stanford‘s computational photography research labs, and have built their own light field camera arrays, and they’re posting instructions on how to build your own, including a method for achieving the effect with just one camera. Check out much more information and some of the science behind the project at FuturePicture.com.
(via MetaFilter Projects)
One of my favorite things to think about is the difficulty of communicating with humans generations from now, or even tens of thousands of years from now. An example: The Office of Civilian Radioactive Waste Management overseeing Yucca Mountain, the proposed Nevada site for disposal of nuclear waste, has been working with artists to develop a warning system that would alert future visitors to the area of the dangers buried in the mountain. From the website, “The monumental challenge is to address how warnings can be coherently conveyed for thousands of years into the future when human society and languages could change radically.” The purpose of the warning sign is “to deter intentional or inadvertent human intrusion or interference at the site and to effectively communicate over the course of the next 10,000 years that the integrity of the site must not be compromised in any way in order to prevent the release of the radiation contained within.” It’s an interesting visual challenge that must not rely on our own cultural biases. Here’s one artist’s response to the challenge, though perhaps it’s too reliant on the 20th century “Radioactive Danger” symbol.
In 1999, the New York Times Magazine ran a six-issue Millenium special, one part of which was an invitation to artists, scientist, and other thinkers, to develop a way of communicating with the future. Jaron Lanier, researcher and scientist, proposed genetically engineering a DNA-coded archive of a year’s worth of the New York Times Magazine and inserting it into the common cockroach’s genome (and the New York Times’ discussion of the idea). Owing to the millions-of-years-long stability of the cockroach genome and the species tenacious ability to survive ice ages, floods, and other earth-altering natural disasters, the cockroach proves to be a perfect candidate. With careful gene splicing techniques, coded DNA could be inserted into unused areas of the cockroach genome, providing a carrier for what could be, if the encoded information expanded beyond the scope of the New York Times Magazine, a living, breathing, self-replicating, everywhere Library of Alexandria (the burning of which illustrates the importance of millenia-long preservation of our academic and cultural knowledge). Under Lanier’s proposal, cockroach reproduction would spread the DNA-coded archive into the every cockroach in New York City in just 14 years. Future humans or other visiting species would hopefully decode this time capsule upon study of the species and human knowledge will have survived across the millenia, regardless of extinction or other disasters.
Weird and ingenious.
While I can’t pretend to know the intricate differences between a jpg and a png, I’ve got an appreciation of the science and mathematics that goes into imaging technology. Every so often a strange new development from MIT’s media lab crosses my radar, as happened a couple weeks ago with Bokode, a barcode hidden and read in out-of-focus areas of a picture which could allow something like physical hyperlinking of data on a webpage to physical objects through photographs made of the image or Nintendo Wii-like interactivity in the classroom or in public areas. Weird stuff with much cool science behind the scenes.
Photosketch, which has since been renamed, is just one such weird academic research project which could radically change the way we interact with photos or, um, decorate our myspace pages. It’s got to be seen to be believed; the video‘s at the top of the post. Basically, you make a couple line drawings with labels, and the software automatically grabs images from the internet and relatively seamlessly creates a new image with all of the specified elements. Sounds crazy. Is crazy.
Here’s a layman’s explanation, but really you should just watch the video. There’s some talk that the whole thing’s a hoax, but the paper announcing the demo came from a couple of researchers associated with related image research and Photosketch was presented at a respected graphics technology symposium. The source code was purportedly available at some point, but it’s disappeared since I first read accounts of the technology. I was hoping to illustrate the post with a picture of me photographing a unicorn in a coral reef while a UFO flew by….
Heather Morton has a roundup of perspectives on the tool.
(first spotted on waxy.org)
I’m always interested in how photography gets used outside of photography circles. Sarah Williams’ and Minna Ninova’s maps of cultural “buzz,” based on geo-tagged photos in the Getty archive, scratch that itch well. Using Elizabeth Currid’s and Sarah Williams’ research data culled from thousands of photos, the maps show heatmaps of what might be considered culturally significant locales. Of course, high-profile media events (the Oscars, for instance) throw some skew into the data, making it difficult for still important but less publicized events to show on the map, but they work as a great illustration of the places where mainstream America casts its gaze. If the accuracy of the data could be controlled for, I’m sure a similarly-created map using flickr’s huge data set would be very interesting. The research paper, “The Geography of Buzz: Art, Culture and the Social Milieu in Los Angeles and New York,” is also available.
(via Lens Culture)