One designer has developed Paragraphica, a generative AI in the form of a camera, which takes snapshots using location data from descriptions of where you are. The project, which is an art show, is interesting.
Between Midjourney, Dalle-E, and many other tools, AI systems for image generation are more than topical. This concept, which makes it possible to create images from raw data, is of interest to many creators. This is the case of the Danish designer Bjorn Karmann. in a tweet Published March 30, 2023, announced the creation of Paragraphicaa very special “camera”.
The idea is simple but clever. By linking location data from its camera to open online APIs, Paragraphica was able to generate a detailed description of the current location and time. But the technical prowess doesn’t stop there, because the device also generates AI-generated images.
To achieve this, Paragraphica uses our location data and other relevant information. For example, at the time of this writing, I am using the program in Paris. By combining my address, weather, time of day, and nearby locations, Paragraphica composes a text detailing representation of the current location and time.
Here, this is the prompt created by the tool: “Morning photo taken at 2 rue Comte, 75003 Paris. The weather is clear with a temperature of 19 degrees. The date is Wednesday, May 31, 2023. There is a nursery, a French restaurant, and a cigarette seller nearby.”
From this AI-generated prompt, the camera produces an image that reflects the way the AI sees that place. If you fail to see results by testing the tool to the end (The fault is in the already significant popularity of the program), the combined examples show satisfactory renderings, all without an existing lens.
Especially as Bjorn Karmann explains on his site“,” The resulting image is not a simple snapshot, but a complex and accurate reflection of where you are. According to the inventor, the images created were able to capture ” Some place-specific vibes and feelings, but in a weird way, because the pictures never seem to be Where the photo was taken.
hard and soft
The tool exists in two different forms. The first, accessible to all interested parties, is a simple virtual camera.
On the other hand, the second form, which is currently only used by Bjorn Karmann, is a physical form of Paragraphica. In this format, the device has a viewfinder, which displays a real-time description of the device’s current position. This viewfinder is accompanied by a trigger, which it is sufficient to press to create an image representation of the description shown, which thus acts as a guide.
To make it feel a little more robust like a camera, Bjørn Karmann made sure to add three thumb control wheels that control AI data and settings.
The first wheel, like the focal length of an optical lens, allows control of the radius (in meters) of the area the camera is searching for places and data. As for the second wheel, it allows you to manage Noise Necessary for the AI-generated image publishing process. And using the third wheel, the user can amplify or decrease the AI’s accuracy in following the prompt. Obviously, the higher the value, the clearer and more accurate the image.
All of these options make it possible to give a more personal view to the photos, as one would do on a conventional camera. But despite this classic approach, Paragraphica, according to Bjorn Karmann, ” A way to understand the world around us that is not limited to visual perception alone “.
Subscribe for free to the Artificial Intelligence Newsletter, Designed by AIs, Verified by Numerama!
“Hardcore beer fanatic. Falls down a lot. Professional coffee fan. Music ninja.”