AI Image Generation Gets A Drag Interface

May 27, 2023 by Donald Papp 15 Comments

AI image generators have gained new tools and techniques for not just creating pictures, but modifying them in consistent and sensible ways, and it seems that every week brings a fascinating new development in this area. One of the latest is Drag Your GAN, presented at SIGGRAPH 2023, and it’s pretty wild.

It provides a point-dragging interface that modifies images based on their implied structure. A picture is worth a thousand words, so this short animation shows what that means. There are plenty more where that came from at the project’s site, so take a few minutes to check it out.

GAN stands for generative adversarial network, a class of machine learning that features prominently in software like image generation; the “adversarial” part comes from the concept of networks pulling results between different goalposts. Drag Your GAN has a GitHub repository where code is expected to be released in June, but in the meantime, you can read the full paper or brush up on the basics of how AI image generators work, as well as see how image generation can be significantly enhanced with an understanding of a 2D image’s implied depth.

Advanced Techniques For Realistic Baking Animations

November 19, 2019 by Lewin Day 6 Comments

Computer graphics have come a long way since the days of Dire Straits and their first computer animated music video in 1985. To move the state of the art forward has taken the labor of countless artists, developers and technicians. Working in just that field, a group from UCLA have developed an advanced system for simulating baking in computer graphics, and the results look absolutely delicious.

We propose a porous thermo-viscoelastoplastic mixture model.

The work is being presented at SIGGRPAH Asia, and being an academic paper, is dense in arcane terminology. To properly simulate baking, the team had to consider a multitude of interdependent processes. There’s heat transfer to consider, the release of carbon dioxide from leavening agents, the browning of dough due to evaporation of water, and all manner of other complicated chemical and physical interactions.

With a model that takes all of these factors into account, the results are amazingly realistic. The team have shown off renders of cookies in the oven, freshly baked loaves of bread being torn apart, and even muffins full of melted chocolate chips.

We imagine it would have been difficult not to work up an appetite during the research process. We’ve seen impressive work from SIGGRAPH before, like this method for printing photorealistic images on 3D surfaces. Video after the break.

Continue reading “Advanced Techniques For Realistic Baking Animations” →

Printing Photorealistic Images On 3D Objects

May 13, 2015 by Theodora Fabio 36 Comments

Hydrographic Printing is a technique of transferring colored inks on a film to the surface of an object. The film is placed on water and activated with a chemical that allows it to adhere to an object being physically pushed onto it. Researchers at Zhejiang University and Columbia University have taken hydrographic printing to the next level (pdf link). In a technical paper to be presented at ACM SIGGRAPH 2015 in August, they explain how they developed a computational method to create complex patterns that are precisely aligned to the object.

Typically, repetitive patterns are used because the object stretches the adhesive film; anything complex would distort during this subjective process. It’s commonly used to decorate car parts, especially rims and grills. If you’ve ever seen a carbon-fiber pattern without the actual fiber, it’s probably been applied with hydrographic printing.

The physical setup for this hack is fairly simple: a vat of water, a linear motor attached to a gripper, and a Kinect. The object is attached to the gripper. The Kinect measures its location and orientation. This data is applied to a 3D-scan of the object along with the desired texture map to be printed onto it. A program creates a virtual simulation of the printing process, outputting a specific pattern onto the film that accounts for the warping inherent to the process. The pattern is then printed onto the film using an ordinary inkjet printer.

The tiger mask is our personal favorite, along with the leopard cat. They illustrate just how complex the surface patterns can get using single or multiple immersions, respectively. This system also accounts for objects of a variety of shapes and sizes, though the researchers admit there is a physical limit to how concave the parts of an object can be. Colors will fade or the film will split if stretched too thin. Texture mapping can now be physically realized in a simple yet effective way, with amazing results.

Continue reading “Printing Photorealistic Images On 3D Objects” →

3-Sweep: Turning 2D Images Into 3D Models

September 12, 2013 by James Hobson 34 Comments

As 3D printing continues to grow, people are developing more and more ways to get 3D models. From the hardware based scanners like the Microsoft Kinect to software based like 123D Catch there are a lot of ways to create a 3D model from a series of images. But what if you could make a 3D model out of a single image? Sound crazy? Maybe not. A team of researchers have created 3-Sweep, an interactive technique for turning objects in 2D images into 3D models that can be manipulated.

To be clear, the recognition of 3D components within a single image is a bit out of reach for computer algorithms alone. But by combining the cognitive abilities of a person with the computational accuracy of a computer they have been able to create a very simple tool for extracting 3D models. This is done by outlining the shape similar to how one might model in a CAD package — once the outline is complete, the algorithm takes over and creates a model.

The software was debuted at Siggraph Asia 2013 and has caused quite a stir on the internet. Watch the fascinating video that demonstrates the software process after the break!

Continue reading “3-Sweep: Turning 2D Images Into 3D Models” →

Filming Light Reflecting Off Objects

May 18, 2013 by Brian Benchoff 13 Comments

light

With high-speed cameras you’re able to see bullets passing through objects, explosions in process, and other high-speed phenomena. Rarely, though, are you able to see what happens when light shines on an object without hundreds of thousands of dollars worth of equipment. A group of researchers at The University of British Columbia are doing just that with hardware that is well within the range of any home tinkerer.

Making videos of light passing through and around objects has been done before (great animated gifs of that here), but the equipment required of previous similar projects cost $300,000 and couldn’t be used outside the controlled environment of a lab. [Matthias] and his team put together a similar system for about $1,000. The only hardware required is an off-the-shelf 3D time of flight camera and a custom driver powering six red laser diodes.

Aside from having a much less expensive setup than the previous experiments in recording the flight of a pulse of light, [Matthias] and his team are also able to take their and record the flight of light in non-labratory settings. They’ll be doing just that at this year’s SIGGRAPH conference, producing videos of light reflecting off attendee-produced objects in just a few minutes. You can check out the video for the project below.

Largest CNC Router Is Controlled By Hand

August 9, 2012 by Brian Benchoff 61 Comments

Fresh from this year’s SIGGRAPH is a very interesting take on the traditional X Y-table based CNC machine from [Alec], [Ilan] and [Frédo] at MIT. They created a computer-controlled CNC router that is theoretically unlimited in size. Instead of a gantry, this router uses a human to move the tool over the work piece and only makes fine corrections to the tool path with the help of a camera and stepper motor.

The entire device is built around a hand held router, with a base that contains a camera, electronics, stepper motors, and a very nice screen for displaying the current tool path. After a few strips of QR code-inspired tape, the camera looks down at the work piece and calculates the small changes the router has to make in order to make the correct shape. All the user needs to do is guide the router along the outline of the part to be cut with a margin of error of a half inch.

You can read the SIGGRAPH paper here (or get the PDF here and not melt [Alec]’s server), or check out the demo video after the break.

Anyone want to build their own?

Continue reading “Largest CNC Router Is Controlled By Hand” →

Improve Your Vision With Computer Generated Glasses

July 16, 2012 by Brian Benchoff 16 Comments

[Vitor Pamplona] sent in a project presented at this years SIGGRAPH. It’s a piece of hardware that corrects vision without the need for lenses. Yep. software-defined eyeglasses now exist, even if the project is a bit bulky for daily wear.

[Vitor] et al came up with two versions of hardware for this project. The first is a dual stack of high-resolution LCD displays, while the second revision is an LCD with a lenticular overlay. With this hardware, the team can change the focal plane of an entire image, or just subsets of an image allowing for customized vision correction for anyone with nearsightedness, farsightedness, astigmatism, presbyopia, and even cataracts.

With plenty of head-mounted augmented reality platforms coming down the pipe such as Google’s Project Glass and a few retina displays, we could see this type of software-defined vision correction being very useful for the 75% of adults who use some form of vision correction. It may just be a small step towards the creation of a real-life VISOR, but we glasses-wearing folk will take what we can get.

You can check out the .PDF of the paper here, or watch the video after the break.

Continue reading “Improve Your Vision With Computer Generated Glasses” →