wbenbihi/hourglasstensorflow
A TensorFlow implementation of Stacked Hourglass Networks for estimating human body keypoint locations from images.

This repository provides a TensorFlow implementation of Stacked Hourglass Networks, a convolutional neural network architecture designed for human pose estimation. The model predicts the locations of body keypoints (such as joints) in images by repeatedly downsampling and upsampling feature maps through hourglass-shaped modules. It is implemented primarily in Jupyter Notebooks and targets researchers and developers working on multi-person pose estimation tasks.