This paper presents a fast and robust multiple object tracking algorithm based on an RGB-D version of the MeanShift tracking algorithm and exploiting RGB-D camera networks when multiple RGB-D sensors are available. The original Mean-Shift algorithm has been improved in three ways. First, a color-depth Joint Probability Density Function is proposed for taking into account both depth and color information. Secondly, we propose an occlusion detection mechanism which can handle long-term occlusions even when objects move fast and unpredictably. Finally, when multiple views are available, we combine the tracking outcomes from all the RGB-D sensors in our network to deal with the identity confusion problem and enhance the overall tracking performance. Experimental results demonstrate that the proposed scheme is robust, realtime and has yielded a marked improvement with respect to the state-of-the-art in terms of tracking quality. As a further contribution, we released our work as open-source in order to provide the best benefit to the wide Computer Vision and Robotics community.

Microsoft Kinect had a key role in the development of consumer depth sensors being the device that brought depth acquisition to the mass market. Despite the success of this sensor, with the introduction of the second generation, Microsoft has completely changed the technology behind the sensor from structured light to Time-Of-Flight. This paper presents a comparison of the data provided by the first and second generation Kinect in order to explain the achievements that have been obtained with the switch of technology. After an accurate analysis of the accuracy of the two sensors under different conditions, two sample applications, i.e., 3D reconstruction and people tracking, are presented and used to compare the performance of the two sensors.

In this work, we describe a novel method for creating 3D models of persons freely moving in front of a consumer depth sensor and we show how they can be used for long-term person re-identification. For overcoming the problem of the different poses a person can assume, we exploit the information provided by skeletal tracking algorithms for warping every point cloud frame to a standard pose in real time. Then, the warped point clouds are merged together to compose the model. Re-identification is performed by matching body shapes in terms of whole point clouds warped to a standard pose with the described method. We compare this technique with a classification method based on a descriptor of skeleton features and with a mixed approach which exploits both skeleton and shape features. We report experiments on two datasets we acquired for RGB-D re-identification which use different skeletal tracking algorithms and which are made publicly available to foster research in this new research branch.

In this paper we propose a novel methodology for people re-identification based on skeletal information. Features are evaluated on the skeleton joints and a highly distinctive and compact feature-based signature is generated for each user by concatenating descriptors of all visible joints. We compared a number of state-of-the-art 2D and 3D feature descriptors to be used with our signature on two newly acquired public datasets for people re-identification with RGB-D sensors. Moreover, we tested our approach against the best re-identification methods in the literature and on a widely used public video surveillance dataset. Our approach proved to be robust to strong illumination changes and occlusions. It achieved very high performance also on low resolution images, overcoming state-of-the-art methods in terms of recognition accuracy and efficiency. These features make our approach particularly suited for mobile robotics.

This paper proposes a very fast and robust multi-people tracking algorithm suitable for mobile platforms equipped with a RGB-D sensor. Our approach features a novel depth-based sub-clustering method explicitly designed for detecting people within groups or near the background and a three-term joint likelihood for limiting drifts and ID switches. Moreover, an online learned appearance classifier is proposed, that robustly specializes on a track while using the other detections as negative examples. Tests have been performed with data acquired from a mobile robot in indoor environments and on a publicly available dataset acquired with three RGB-D sensors and results have been evaluated with the CLEAR MOT metrics. Our method reaches near state of the art performance and very high frame rates in our distributed ROS-based CPU implementation.