Excellent talk, thank you Michael and the rest of the MPI / ELLIS / Cyber Valley team!

Thanks for your great job. If I reconstruct 3d model from one 2d image using smpl-x, how can I use this to random shape. I have tryed it, it seams not work. It gives error "line 366, in execute for key_block in obj.data.shape_keys.key_blocks: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AttributeError: 'NoneType' object has no attribute 'key_blocks'"

👏🏽👏🏽👏🏽

this just looks like random hand motion tbh

Thanks for this solution

I try to test code but while running code in colab. What should i upload in colab as input image to see output? By the way, Nice work!!

Awesome. Objective of life: (1st) work with human mesh modelling and graphics. (2nd) Have 1% of Michael's research skills.

Very much enjoyed the history. Thanks for all your work and making knowledge availble to general audience like me.

I'm sorry did you say something?

how can I get the username and password

fantastic lecture, thank you. I wonder if the next iteration of smpl will be something akin to segment anthing. Which processes millions of masks given a prompt to find the most likely match.

Hello guys, but may be do you have video how to install code if you a not programmer? Because i 3d character animator in maya, and very interesting to try! OR maybe detailed instructions? Thank you!

Are the Unreal environments that were used to record the scenes available for research purposes? Or do I need to buy them on the Unreal marketplace?

THE FUTURE

Michael you are an extraordinary person!

Hello bro please tell me how I can annotate it manually on a human image

Very interesting video. I also find it interesting how the animated model (3:06) is leaning forward and not having an vertically aligned posture as the effects of f.e. locked knees, forward tilted hips and the effects on the spinal curves determine the breathing mechanism. In walking (and standing), kinethic energy can be transmitted troughout the body to the diaphragm and ribcage, affecting its tone and activity, too. Hand and arm movement influence the breathing muscles, too. This was just a slight insight but I wonder how biomechanical aspects of this kind are thought about, too?

Hello, is there any way that I can set the collider on the SMPL-X character? Appreciate it!

Fun!

Please make a tutorial

Michael, are you aware if anyone has retrained DECA on synthetic data?

Hello, I was trying to transfer smpl-x action data in UE to Metahuman with IK Retarget , but I found that hand action is distorted and that's why.

To me it looks like it totally loses the identity compared to DECA. Also, the expressions also look exaggerated and not like in the original image. It would have been interesting to see the rendered deformed mesh with the extracted textures.

Does this add on have any scripting documentation? Wanna load our params into the system.

Good job. Most of the camera views of the source videos are from the front, but if the camera is mostly overhead will the animated 3D reconstruction be as correct?

LOL

Thank you for the amazing work! I’m wondering if there’s a way to apply this code to create a lively animated face, similar to Apple's Memoji, to replace man head in a video?.

😍😍😍👏🙌 love this

Can imagine how much time to prepare such realistic dataset!

wow so cool

awsome!

I too wonder how it would perform on poses with rapid acceleration where the center of mass is not necessarily supported by a center of pressure directly below. For example, an athlete making a sharp hairpin turn would have significant lean, but because of the acceleration AND gravity, the center of mass would be supported despite the lean.

I’m convinced that this is how we get there! Of course your team is the one to do it.

Great job on understanding the importance of maintaining a stable body and its connection to slow movements like Yoga. I'm curious if you've explored its effectiveness in higher-intensity activities such as fast walking, running, or jumping. Considering deformable surfaces adds another layer to consider, especially when the ground is not always a non-deformed surface, like when people play on grass, for example. It would be an excellent addition to your original concept and research model. Keep up the fantastic work!

Professor, can U tell me if this dataset can be used for camera synchronization or not ?

In fact, when someone gets weight the skeleton also changes a little bit anatomically. Also, the visual joints are not always the same as the anatomical joints/skeleton. So it changes a lot when get or lose weight, I think.

Thank you for sharing the presentation, I have a couple of questions: Firstly the global position and orientation is obtained through camera motion estimation or directly for each tracking object? Secondly, the re-identification is based on the shape parameters of SMPL models of humans or others? If yes, what if the people's body shape is similar to each other. Finally, It only works for humans or other types of objects?

wow. now only need to differntiate between different weight, weft and material of fabrics, they all drape different

is there any tutorial on how to implement this on a persnal image? i have no background in the feild but i am curious about it. i'd like to have a first hand experience.

1. Can the model be exported without cropping to the roi box? 2. What can be done to improve the temporal stability / shakiness? Thank you!

How to do it ? Full tutorial please

Is there any way to create these animations online to use as video for your website?🤔

Code is filled with bugs and errors

Is it possible to get a model with texture?

Truly amazing, would very much like to collaborate with you at some point in my career.