Piercarlo Dondi

Scaling up to National/Regional Urban Extent Mapping Using Landsat Data

This paper describes a methodology to extract a consistent human settlement extent layer using Landsat data and its implementation in the Google Earth Engine platform. The approach allows the extraction of human settlement extents by means of the existing Landsat 5 and 7 data sets, allowing to check their evolution at 30-m spatial resolution. Since human settlements are the main proxy to people geographical distribution and to building locations, this layer may serve as a mean to disaggregate people/building counts at the regional/national level. The approach is tested in several parts of the world against existing ground truth data at the same spatial resolution in Brazil and China, as well as against extents manually extracted from VHR data in three different geographical areas: 1) Brazil; 2) South East China; and 3) Indonesia.

Development of gesture-based human–computer interaction applications by fusion of depth and colour video streams

Hand detection and gesture recognition are two of the most studied topics in human–computer interaction (HCI). The increasing availability of sensors able to provide real-time depth measurements, such as time-of-flight cameras or the more recent Kinect, has helped researchers to find more and more efficient solutions for these issues. With the main aim to implement effective gesture-based interaction systems, this study presents an approach to hand detection and tracking that exploits two different video streams: the depth one and the colour one. Both hand and gesture recognition are based only on geometrical and colour constraints, and no learning phase is needed. The use of a Kalman filter to track hands guarantees system robustness also in presence of many persons in the scene. The entire procedure is designed to maintain a low computational cost and is optimised to efficiently execute HCI tasks. As use cases two common applications are described: a virtual keyboard and a three-dimensional object manipulation virtual environment. These applications have been tested with a representative sample of non-trained users to assess the usability and flexibility of the system.

Evaluation of a Foreground Segmentation Algorithm for 3D Camera Sensors

Our interest is focusing on the innovative time-of-flight (TOF) cameras. In this paper we present our approach to foreground segmentation, based on smart-seeded region growing. The seeding strategy makes use of the characteristic intensity signal provided by TOF cameras, and growing is proved by experimental measurements to produce a pixel-wise segmentation of 82%-92% quality. Compared to background subtraction, our approach uses more explicitly the unique capacity of TOF cameras to isolate foreground objects on the basis of their distance. Our work will find an application in gate monitoring and passage surveillance.

Automatic Analysis of UV-Induced Fluorescence Imagery of Historical Violins

In recent years, UV-induced fluorescence (UVIFL) photography has proven to be very effective when studying the surface of historical musical instruments, such as violins. This technique makes it possible to highlight superficial details not clearly perceptible with visible light (e.g., retouchings, superficial distribution of varnishes, or wear). The data retrieved are also an important guide for further noninvasive spectroscopic analyses used when the chemical composition of the surface needs to be investigated. However, UVIFL imagery interpretation of a historical violin is no trivial task. In fact, constant playing and the multiple restorations over the centuries have produced very complex surfaces. This work presents an automatic tool designed to facilitate this kind of analysis. Using a quantized histogram in HSV color space, the distribution of the main fluorescence colors on an instrument’s surface can be highlighted, recurrence of the same color in different areas of the same violin can be detected, or different violins can be compared. UVIFL images of seven Stradivarius violins kept in the Museo del Violino in Cremona, Italy, were used as a test set. The results achieved endorse the validity of the proposed approach.

An Interactive Tool for Speed up the Analysis of UV Images of Stradivari Violins

UV fluorescence photography is widely use in the study of artworks, in particular for the analysis of historical musical instruments. This technique allows seeing important details which cannot be observed with visible light, such as retouching, different paints coats or worn areas. The complexity of the interpretation of the surface of a violin is proportional to its state of preservation: more alterations correspond to a more wide range of colors. We designed an interactive tool able to help the scientist to understand the composition of the surface and in particular the distribution of the colors on the entire instrument, avoiding perception illusion. The result is achieved using a quantized histogram in HSV color space. The tests were performed on UV imagery of the Stradivari violins collection stored by “Museo del Violino” in Cremona.

Multisubjects Tracking by Time-of-Flight Camera

Time-of-Flight cameras are the state of art sensors for a fast detection of depth data in a scene. This kind of sensors can be very useful for tracking, in particular in indoor ambient, since, using light in near-infrared spectrum, they are less affected by abrupt change in illumination. In this paper we propose a new method for the tracking of multiple subjects based on Kalman filter. The first step of our solution is a ToF based foreground segmentation, that retrieves all significant clusters in the scene, followed by a robust tracking system able to correctly handle occlusions and possible merging between clusters.

Human-computer interaction through time-of-flight and RGB cameras

The number of systems exploiting Time-of-Flight (ToF) cameras for gesture recognition has greatly increased in the last years, confirming a very positive trend of this technology within the field of Human-Computer Interaction. In this work we present a new kind of application for the interaction with a virtual keyboard which is based on the use of an ordinary RGB webcam and a ToF camera. Our approach can be subdivided into two steps: firstly a segmentation of the entire body of the user is achieved exploiting only the ToF data; then the extraction of hands and head is obtained applying color information on the retrieved clusters. The final tracking step, based on the Kalman filter, is able to recognize the chosen hand also in presence of a second hand or the head. Tests, carried out with users of different ages, showed interesting results and a quick learning curve.

Multimodal workflow for the creation of interactive presentations of 360 spin images of historical violins

The adoption of multimedia and multimodal applications inside museums and exhibitions is becoming a common practice. These installations proved to be particularly effective to attract visitors, especially the younger ones, and teach them complex information about the exposed artworks. A similar approach can be useful also to explain scientific analyses conducted on artworks, that commonly involve the use of complex analytical techniques, difficult to understand for inexpert people. In this work, we describe a multimodal workflow for the creation of interactive presentations of 360 spin images of historical violins. The workflow involves the acquisition, classification and visualization of the data. In particular, an ad hoc photographic set was built to achieve a fast acquisition of images both under visible and UV illumination, with the aim to study the surface of the instruments. Acquired images are classified and labeled using UVAnalyzer, an interactive application that supplies a set of tools for the analysis of UV fluorescence (UVF) images. Finally, KVN (Kinect Violin Navigator), a Kinect-based application, provides comfortable visualization and navigation within the data. During the development, we adopted a user-driven approach: user studies and suggestions from experts in UVF analysis were taken into account to improve the usability of the various steps of the workflow; then, a qualitative evaluation of the produced presentation was conducted on 22 volunteers. As a test set, we used images of violins exposed in the “Museo del Violino” in Cremona (Italy).

Real-Time Foreground Segmentation with Kinect Sensor

In the last years, economic multichannel sensors became very widespread. The most known of these devices is certainly the Microsoft Kinect, able to provide at the same time a color image and a depth map of the scene. However Kinect focuses specifically on human-computer interaction, so the SDK supplied with the sensors allows to achieve an efficient detection of foreground people but not of generic objects. This paper presents an alternative and more general solution for the foreground segmentation and a comparison with the standard background subtraction algorithm of Kinect. The proposed algorithm is a porting of a previous one that works on a Time-of-Flight camera, based on a combination of a Otsu thresholding and a region growing. The new implementation exploits the particular characteristic of Kinect sensor to achieve a fast and precise result.

Automatic selection of regions of interest in a video by a depth-color image matting

The automatic detection of regions of interest in a video is fundamental for a fast generation of many ground truth images. In this paper we introduce a new solution for selecting regions of interest based on an automatic image matting method. Image matting is a set of techniques designed to obtain a precise separation of background and foreground in image or video sequences. Basically all the matting approaches need a direct human interaction, there are only few total automatic solutions. To achieve this goal we combine two different video streams: the color one and the depth one. In particular, we use an automatic depth based segmentation to substitute the human input in the Soft Scissors, one of the most precise matting algorithm. The overall efficiency is achieved using the Nvidia CUDA architecture to execute the most computational intensive sections of algorithm. The result of the matting can be used as a ground truth for successive elaborations.