Author: Rauf Aliev

Liza at work. Mastering charcoal. I’ll post a photo of the charcoal piece she used to draw this in the comments

The service allows you to upload your photo and get a 3D model. You can even use it with a photo of the president.

http://cvl-demos.cs.nott.ac.uk/vrn/

Tried uploading myself – it turned out pretty well.

Here are the details:

Here are the sources on GitHub:

Here is another work by one of the authors – already with human figure modeling:

Konstantin Kharitonov you seem to understand this. Why isn’t there a tool for creating a 3D model of a person in motion (by taking photos of them from multiple angles, asking them to make various facial expressions, possibly by projecting a laser grid on them), then process what was captured, and by combining it with some universal knowledge base about human anatomy, achieve a 3D model of a specific face, which can then depict any desired facial expression on a computer just because a) muscles work the same way in all people b) the model knows the ranges in which certain muscles move c) the model knows how the skin, eyes, mouth, and ears look when certain muscles are tensed or stretched.

After all, no one limits us in the number of photos taken from different angles and under different lighting conditions. So even if there are a million photos, it’s no problem, we’ll wait. But then, the system could extract the photo from the needed perspective under the required lighting to render realistic expressions of a specific face for any artist’s requirement.

I imagine this as a framework for 3D designers, allowing them to simply specify any facial expression, but also to alter the face itself (e.g., widen the eyes or protrude the nose). With a 3D model from a camera, import these facial features from a “scanner”, as well as the skin texture for more accurate rendering.

Is there such software?

For those interested in research in the field of video processing – head here

Lots of fascinating video reports from Berkeley labs

It turns out that the sugar content in horseradish ranges from 6% to 13%, while in radish it ranges from 1.5% to 6.4%. In other words, it is generally possible to find horseradish that is not sweeter than radish, but more often than not, it is sweeter.

Hmm…

https://github.com/aeroflotsrc/webapp

“Murder of crows” and “Parliament of owls” are the best. People say “an embarrassment of pandas” is a great addition. Embarrassment!

Excellent channel about mathematics and a great storyteller (Matt Parker). In the first video below, he talks about a cool paradox with multidimensional spheres (don’t worry, it’s all simple). In the second – about the mysterious number 10958 (+ solution to this number’s problem in a separate video)

https://www.youtube.com/watch?v=mceaM2_zQd8

https://www.youtube.com/watch?v=-ruC5A9EzzE

At every corner and along the streets of Richmond’s university campus, there are many cameras. There are thousands of them here. At this corner, there are eight. It seems the goal was the complete absence of blind spots and, judging by the infrared lighting, they operate even at night.

The same pianist, many years apart. Both videos are masterpieces

https://www.youtube.com/watch?v=ci4CQlRSBEs

https://www.youtube.com/watch?time_continue=277&v=gYtybgToH2Q