You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Pipeline to accurately and efficiently compute the shape and orientation (pose) of a car given its RGB image by using keypoints generated from an hourglass network after optimising the estimates using the Ceres solver
tf.keras implementation of "Toward fast and accurate human pose estimation via soft-gated skip connections" by Bulat et al. and "Stacked Hourglass Networks for Human Pose Estimation" by Newell et al.
This repository about the project of Computer Vision course at the Sapienza University of Rome: Single-Person 2D Joints Estimation using Convolutional Neural Network, Stacked Hourglass Network, etc.