live-pose

Intel RealSense Depth Camera compatible Python package for live 6 DOF pose estimation.
For Jetson device use Jetpose.

Table of Contents

• Installation
• Preparation
• Usage

Installation

1. Clone the repository:

   git clone https://github.com/prajwalgt/live-pose-FastSAM.git
   cd live-pose

Preparation

Docker Build

1. Build the Docker container:

   cd docker
   docker build --network host -t foundationpose .

CLIP should be installed when building the image, please check.

2. Install Weights:
   • Download the weights from this link and place them under live-pose-FastSAM/FoundationPose/weights.

Usage

Running the Container

1. Run the container:

   bash docker/run_container.sh

   note: To run on windows install Cygwin and execute ./docker/run_container_win.sh

Building packages

1. Build:

   CMAKE_PREFIX_PATH=$CONDA_PREFIX/lib/python3.9/site-packages/pybind11/share/cmake/pybind11 bash build.bash

Running the Model

1. Run the live pose estimation:

   bash run_live.sh

   note: To run on windows install Cygwin and execute ./run_live.sh

2. Locate the .obj file:
   Note: For novel object you can use Object Recustruction Framework
   

3. Masking:
   here select the Boundry points of object in first frame