In the aftermath of an earthquake, conducting rapid structural safety assessments is essential. A Deep Learning-based damage detector capable of automatically analyzing videos from Unmanned Aircraft Systems (UAS) surveys would be highly beneficial for this purpose. Despite significant advancements in object detection using Deep Convolutional Neural Networks (DCNNs), developing an effective post-earthquake damage detector remains challenging due to the scarcity of large, annotated image datasets. In this work, we present a system to create a large number of images where artificial damage instances are applied to real-world three-dimensional (3D) models of buildings and bridges. We defined such images as semi-synthetic. The proposed method relies on the definition, made by human experts, of meta-annotations from which a variety of damage instances can be generated in a controlled way. Semi-synthetic images are designed to augment real-world datasets, enhancing the training process of a DCNN-based damage detector. This semi-synthetic image augmentation can be iteratively refined to target the most critical cases. Experiments conducted on the ‘Image Database for Earthquake damage Annotation’ (IDEA) dataset shown that a detector trained on a combination of real and semi-synthetic images performs better than one trained on real images alone. A damage detector trained using the proposed strategy was then incorporated into a system that analyzes and tracks multiple damage instances in UAS-acquired videos, generating concise summaries of the findings. The effectiveness of the system was validated by the analysis of post-earthquake UAS videos and the production of reports that were reviewed by structural engineering experts.

Robots, very widespread in industrial settings, are now used for many different purposes. Even if several tasks can be fully automated, the presence of a human operator is often needed. However, direct robot control is generally complex, and it is thus important to explore alternative and comfortable ways of communication. In this paper, we propose a hybrid approach to human-robot interaction that exploits both manual and gaze-based input. While simple, our solution goes in the direction of enhancing the control capabilities of robots by human operators.

The study of historical documents is an important and complex branch of document analysis. Ancient manuscripts are often altered by time or environmental conditions and the availability of large data sets is not always guaranteed. In this work, we propose a new solution for the attribution of short historical documents based on both allographic and texture features. Paleographic data were used as reference during the development. Tests were conducted on Antonio Stradivari's relics, a collection of technical drawings and artifacts of the famous violin maker, currently held in "Museo del Violino" in Cremona (Italy), that contains his working notes. This data set is very complex, it contains only short writings (few text lines or words), faded or damaged areas, different supports (wood or paper), and various annotations added by the various owners of this collection during the centuries. Experimental results are promising showing an accuracy greater than 90% using short texts both as training and as target.

In recent years, the technological improvements of consumer robots, in terms of processing capacity and sensors, are enabling an ever-increasing number of researchers to quickly develop both scale prototypes and alternative low cost solutions. In these contexts, a critical aspect is the design of ad-hoc algorithms according to the features of the available hardware. This paper proposes a prototype of an autonomous robot for mapping unknown environments and recognizing target objects. During the setup phase one or more target objects are shown to the RGB camera of the robot which, for each of them, extracts and stores a set of A-KAZE features. Afterwards, the robot adopts the ultrasonic distance measurement and the RGB stream to map the whole environment and search a set of A-KAZE features matchable with those previously acquired. The paper also reports both preliminary tests carried out on a reference indoor environment and a case study performed in an outdoor one that validate the proposed system.

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.

PRAM CREW parallel algorithms requiring logarithmic time and a linear number of processors exist for sliding (LZ1) and static dictionary compression. On the other hand, LZ2 compression seems hard to parallelize. Both adaptive methods work with prefix dictionaries, that is, all prefixes of a dictionary element are dictionary elements.Therefore, it is reasonable to use prefix dictionaries also for the static method. A left to right semi-greedy approach exists to compute an optimal parsing of a string with a prefix static dictionary. The left to right greedy approach is enough to achieve optimal compression with a sliding dictionary since such dictionary is both prefix and suffix. We assume the window is bounded by a constant. With the practical assumption that the dictionary elements have constant length we present PRAM EREW algorithms for sliding and static dictionary compression still requiring logarithmic time and a linear number of processors. A PRAM EREW decoder for static dictionary compression can be easily designed with a linear number of processors and logarithmic time. A work-optimal logarithmic time PRAM EREW decoder exists for sliding dictionary compression when the window has constant length. The simplest model for parallel computation is an array of processors with distributed memory and no interconnections, therefore, no communication cost. An approximation scheme to optimal compression with prefix static dictionaries was designed running with the same complexity of the previous algorithms on such model. It was presented for a massively parallel architecture but in virtue of its scalability it can be implemented on a small scale system as well. We describe such approach and extend it to the sliding dictionary method. The approximation scheme for sliding dictionaries is suitable for small scale systems but due to its adaptiveness it is practical for a large scale system when the file size is large. A two-dimensional extension of the sliding dictionary method to lossless compression of bi-level images, called block matching, is also discussed. We designed a parallel implementation of such heuristic on a constant size array of processors and experimented it with up to 32 processors of a 256 Intel Xeon 3.06 GHz processors machine (avogadro.cilea.it) on a test set of large topographic images. We achieved the expected speed-up, obtaining parallel compression and decompression about twenty-five times faster than the sequential ones.

The method proposed in this paper for two-dimensional object recognition relies on the idea of a structural coding of an object at varying levels of resolution. A tree structure describes the evolution of the contour at increasing levels of detail. Each tree node represents a contour segment through a set of attributes to provide a richer description of the image shape. In addition to the attributes, a weight is associated to each descriptor indicating the significance of the corresponding primitives. An approach for 2D object recognition based on a matching of the attributed tree representation of the candidates with that of the model is proposed.

Among the various proposal advanced to build massive parallel systems in which the number of computing units ranges in the thousands, hierarchical topologies share a number of interesting properties. The authors review these architectures and their applicability and reliability, with particular attention to connections complexity and the ability to exchange messages. The usual assumption of the multiple instruction multiple data (MIMD) computational paradigm is as follows: autonomous but cooperating tasks execute on different processing units in the system. The overall complexity of the systems is measured with the analysis of the diameter and of the increasing law that states the number of interconnections against the number of nodes in the system. The various architectures are compared in terms of links load and average internode distance.>

Hierarchical multi-microcomputer systems are reviewed, with particular attention given to the complexity of the interconnection structure and to the ability to exchange messages. The assumption of the multiple-instruction multiple-data (MIMD) computational paradigm is followed; autonomous but cooperating tasks execute on different processing units in the system. The overall complexity of the systems is measured with the analysis of the diameter and of the increasing law, which states the number of interconnections against the number of nodes in the system. The various architectures are compared in terms of links load and average internode distance under the uniform reference model of message distribution and of request independence assumption.>

An approach to object recognition in an industrial environment is described. With the hypothesis that the universe of objects is limited and that the attitudes in space of any object are constrained by the working environment to a few possibilities, a method is obtained that determines minimal descriptions for each model/object pair. These descriptions, which are discriminant of each model, are used by the recognition strategy planner. It automatically determines which descriptions can be conveniently used to identify the object(s) looked for and in what order. Uncertainties due to overlapping are signaled and tentative classifications listed.>

The authors describe the advantages of the hierarchical implementation of the Hough transform with respect to the labelled one. They demonstrate that the hierarchical Hough is faster than the labelled Hough if simple conditions are satisfied. They then discuss the automatic construction of the hierarchical models for the objects belonging to a completely known set. These models can be seen as oriented graphs whose nodes represent local features or patterns which are subparts of the object silhouette. Some experimental results are presented.>