hi how did you actually use data sets? are they a series of pictures of you doing the t pose? or can the data sets be a live extraction of keypoints while you are doing the t pose while moving around??
The pipeline is composed of two neural network: (1) a pose estimation model which generate body keypoints from images, and (2) a pose classification that takes keypoints as input. The dataset to train the pose classification (e.g. detect the T pose) is thus keypoints generated from a pose estimation model. I've created this keypoints dataset by recording videos of myself in a specific pose.
Hello Arthur nice project!
Can you please write the versions of Jetpack and other dependencies you used as I m encountering some issues?
hi how did you actually use data sets? are they a series of pictures of you doing the t pose? or can the data sets be a live extraction of keypoints while you are doing the t pose while moving around??
The pipeline is composed of two neural network: (1) a pose estimation model which generate body keypoints from images, and (2) a pose classification that takes keypoints as input. The dataset to train the pose classification (e.g. detect the T pose) is thus keypoints generated from a pose estimation model. I've created this keypoints dataset by recording videos of myself in a specific pose.
Hi again Arthur, I am having some problems. Can you help me ?
Maybe, what's your issue?