Shulou(Shulou.com)11/24 Report--

What happens if you connect Stable Diffusion to the robot?

Tell it what you want it to draw, and then draw it one stroke at a time.

While talking and laughing with the human engineer next to him, he skillfully drew a cat with a simple pen on the whiteboard with a marker.

Finally, I didn't forget to sign my name "Ameca". Then I raised my pen hand and continued to examine the work I had just finished with my eyes.

When Ameca asked human beings if they were satisfied with his work, the engineer replied bluntly, "this painting is a little short."

And then the high energy came.

Ameca choked back angrily:

If you don't like my paintings, you probably just don't understand art.

The robot Ameca, from the British company Engineered Arts, has been known for its frighteningly realistic expression once before.

Now it get to the new ability to draw with a pen, and amazed a lot of netizens.

Gently raise the pen after each stroke, just like a real sketch artist, this little detail is so cool.

How does a robot draw with a pen? Exactly how Ameca learned to draw can be explained by itself:

I use the open source Stable Diffusion to skeletonize and vectorize the generated images, and then convert the converted tracks into actions and execute them.

In this way, Ameca's body movements and facial expressions are driven by Engineered Arts's own technology, and the language ability comes from GPT-3 (yes, not 3.5), plus Stable Diffusion makes up a robot painter.

Some netizens said that this ability is very useful in education, and it is much more interesting for students to use a robot than to stare at a computer screen.

Engineered Arts self-developed technology includes robot operating system Tritium and robot production platform Mesmer.

First of all, 36 cameras are used to capture a large amount of human facial data, including bone structure, skin texture and expression.

Do 3D modeling of these data, and 3D print out the silicone skin of the robot.

The Tritium operating system connects hardware, software and the cloud to drive every component of the robot, affecting not only the face, but also the different "muscles" of the head, neck and limbs.

Finally achieve a more lifelike expression.

Coupled with the language model GPT-3 and voice technology, Ameca has initially acquired the ability to interact with human beings.

In addition to painting this time, Engineered Arts also released another video of humans discussing music with Ameca, perhaps suggesting that this is the next challenge.

Ameca says he likes jazz classical and electronic music, and humans have given him the problem of "what are the first lyrics if you want to combine these three styles to write a song?"

Ameca gave an answer after a little thought.

In addition, Ameca is also engaged in an activity to interact with netizens.

If you have any questions to ask it, you can go to the original video comment area to leave a message, and the next video may be Ameca to answer your question.

Reference link:

[1] https://www.youtube.com/watch?v=_1OA_9_wW70

[2] https://www.youtube.com/watch?v=L32BRcvnWRU

[3] https://www.engineeredarts.co.uk/mesmer-creation/

This article is from the official account of Wechat: quantum bit (ID:QbitAI), author: Mengchen

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.