Explore projects

This repo contains a ROS package that computes 3D poses given an array of 2D (image frame) OpenPose detections

C++ wrapper for the mediapipe library capable of loading most graphs.

Multi-modal ROS-based people detection and tracking framework for mobile robots developed within the context of the EU FP7 project SPENCER.

ROS node to extract cluster from Pointclouds and publish them as Spencer Person detections.

Generic navigation, costmap and planner modules to be used as a base for robot specific navigation modules.

ROS package to estimate the depth of the mediapipe skeleton from known average human body distances.

ROS IRI module to handle the head tracking using information from any skeleton provided by the Mediapipe ROS node.

ROS node to transform the YOLO detections into Spencer people tracking detections and estimate their depth using the LIDAR info.

ROS node to transform the YOLO detections into Spencer people tracking detections.

This project uses the resulting segmentation masks from YOLOV8 along a LiDAR to create semantic landmarks that are to be fed to a estimation backend.

Publishes a tf transform which can dynamically be changed using parameters (rqt_reconfigure).

ROS node to get multiple skeletons using YOLO and mediapipe.

Filters Twist messages using LaserScan message information, not allowing to run into obstacles.

Global planner for the move_base framework using the opendrive road description to generate global paths.

Simple local planner for an ackermann platform architecture.

ROS package to generate the input required by the Spencer people tracking stack.

Simple node to detect blobs in images using the HSV color space