Data Sets

Annotated data sets used in our papers:

• Annotated 2D range data sets used in the papers "Place-Dependent People Tracking", IJRR, 30(3) 2011 and ISRR 2009 and "Better Models For People Tracking", ICRA 2011
• Annotated RGB-D data set from our two IROS 2011 papers "People detection in RGB-D data" and "People Tracking in RGB-D Data With On-line Boosted Target Models"
• Annotated 3D range data set from our AAAI 2010 paper "A Layered Approach to People Detection in 3D Range Data" and "Tracking People in 3D Using a Bottom-Up Top-Down Detector", ICRA 2011
• SRL Human Attributes dataset from our ICRA 2015 paper "Real-Time Full-Body Human Gender Recognition in (RGB)-D Data"

Octave/Matlab Tutorial

A comprehensive 121-slide tutorial on Octave and Matlab for beginners and intermediate to advanced users. It covers variables, data types, matrices, plotting, programming, functions and scripts, file I/O, including code examples and practical hints.

• Octave-Matlab-Tutorial.pdf
• Octave-Matlab-Tutorial.key (Keynote format)
• Octave-Matlab-Tutorial.ppt (older version, Powerpoint format)

librobotics

librobotics is a small but useful library of Matlab/Octave functions frequently used in robotics, in particular for plotting. It offers functions to plot different robots, transforms, reference frames, scalable text, error ellipses, or compound operations. librobotics is compatible with both, Octave and Matlab, and sports a full reference documentation.

• librobotics.zip (100 kbytes)
• Reference.pdf

Download, unpack, and add it to your Octave/Matlab path: addpath('librobotics');.
To get help, type help command.

FLIRTLib: Generalized features for 2D range data

FLIRTLib is a library that implements the Fast Laser Interest Region Transform (FLIRT) introduced by Tipaldi and Arras. The library implements four different multi-scale feature detectors and two feature descriptors for 2D range data. It is written in C++ and comes with an API reference (using Doxygen) and a set of example binaries to visualize the detector results and the descriptors, as well as to perform scan-to-scan matching.

• FLIRTLib documentation and download is available here.

People2D: Real-time people detection in 2D range data

People2D is an efficient implementation of the people detector for 2D range data proposed in the ICRA 2008 paper "Human Detection using Multimodal and Multidimensional Features". It is written in C++, includes training and detection source code and example files. The code is written in a way that it is very easy to modify or hack. NO background subtraction is needed with this method. The detector runs at 2 kHz on a modern PC.

• People2D documentation and download is available here.