maghoumi / deepnag

Deep Non-Adversarial Gesture Generation

Hi there 👋

My name is Mehran Maghoumi and I'm a senior deep learning engineer at NVIDIA. My primary area of work is parking space perception using surround camera setups for autonomous driving. I also hold a Ph.D. degree in computer sceince from the University of Central Florida. Feel free to checkout my full profile on my homepage.

What's all this? 🤔

Below is the list of my open source projects that ✨ I'm the most proud of ✨. I've worked on these either during my spare time or as a part of my Ph.D. dissertation. Countless hours of my time have gone into the development of each one, and nothing makes me happier than seeing people use them in their projects.

If you see something you like, please consider ⭐ starring ⭐ the repo. It gives me a better idea of where to focus my efforts!

Happy browsing! 💥

deepnag's Issues

Model is not capturing the hand poses, please guide how can we modify the code.

Kinect v2

Hello can this project use by kinect v2?

Input shape of Soft DTW

Hi, I just reach your paper, Could I have a question the what is the input shape of x and y in this line

DeepNAG/models/hausdorff_distance.py

Line 140 in 1f6a16a

def _pdist_func(self, x, y, mask):

Because I also try Soft DTW, but it uses too much memory, what are the batch size and GPU resources you use in this project?

Thank you!