Celmar Guimarães da Silva

Analysts may use matrix -based visualizations (such as heatmaps) to reveal patterns of a dataset with the help of reordering algorithms that permute matrix rows and columns properly. One of these algorithms is Polar Sort, a pattern -focused reordering method that uses a multidimensional projection technique – Classical MDS – to reveal Band and Circumplex patterns in reorderable matrices. Despite its good reordering results regarding the mentioned patterns, Polar sort is not scalable due to Classical MDS’ asymptotic time complexity (O( n³) for an input matrix with size n × n). In this paper, we propose a new version of this algorithm, in which we replace Classical MDS with FastMap, a method with asymptotic time complexity O(n). The new algorithm (Polar Sort with Fastmap, or PSF for short) permutes rows and columns according to their bidimensional projections and uses a barycenter -based ordering identical to Polar Sort’s approach. The results of an experiment indicate that PSF maintained the output quality of Polar Sort regarding minimal span loss function, Moore stress, and circular correlation when reordering synthetic matrices. Besides, PSF’s asymptotic time complexity is O(n log n). This complexity is coherent with our experiment results, which point out that PSF had lower execution time than other compared methods. We also show some examples in which real -world matrices reordered by PSF revealed patterns similar to Band and Circumplex.

Damiano Oriti, Andrea Sanna, Francesco De Pace, Federico Manuri

Augmented reality (AR) and virtual reality (VR) applications can take advantage of efficient digitalization of real objects as reconstructed elements can allow users a better connection between real and virtual worlds tha n using pre-set 3D CAD models. Technology advances contribute to the spread of AR and VR technologies, which are always more diffuse and popular. On the other hand, the design and implementation of virtual and extended worlds is still an open problem; affo rdable and robust solutions to support 3D object digitalization is still missing. This work proposes a reconstruction system that allows users to receive a 3D CAD model starting from a single image of the object to be digitalized and reconstructed. A smartphone can be used to take a photo of the object under analysis and a remote server performs the reconstruction process by exploiting a pipeline of three Deep Learning methods. Accuracy and robustness of the system have been assessed by several experiments and the main outcomes show how the proposed solution has a comparable accuracy (chamfer distance) with the state -of-the-art methods for 3D object reconstruction.

Douglas Omwenga Nyabuga, Guohua Liu

Hyperspectral imaging (HSI) classification has recently become a field of interest in the remote sensing (RS) community. However, such data contain multidimensional dynam ic features that make it difficult for precise identification. Also, it covers structurally nonlinear affinity within the gathered spectral bands and the related materials. To systematically facilitate the HSI categorization, we propose a spectral-spatial classification of HSI data using a 3D -2D convolutional neural network and inception network to extract and learn the in -depth spectral-spatial feature vectors. We first applied the principal component analysis (PCA) on the entire HSI image to reduce the or iginal space dimensionality. Second, the exploitation of the spatial hyperspectral i nput features contiguous information by 2 -D CNN. Besides, we used 3 -D CNN without relying on any preprocessing to extract deep spectral -spatial fused features efficiently. The learned spectral-spatial characteristics are concatenated and fed to the incepti on network layer for joint spectral -spatial learning. Furthermore, we learned and achieved the correct classification with a softmax regression classifier. Finally, we eval uated our model performance on different training set sizes of two hyperspectral rem ote sensing data sets (HSRSI), namely Botswana (BT) and Kennedy Space Center (KSC), and compared the experimental results with deep learning -based and state-of-the-art (SOT A) classification methods. The experiment results show that our model provides competitive classification results with state -of-the-art techniques, demonstrating the considerable potential for HSRSI classification.

Hanna Koskinen, Susanna Aromaa, Vladimir Goriachev

Our transport system is currently undergoing fundamental change due to increasing use of automation. New automation solutions are introduced in all sectors of transportation, for example automated metros and autonomous ships are in the visions of technology developers. There are many reasons for this ongoing trend of higher use of automation such as demands for sustainability and efficiency to mention some. In this paper, we present a research and development effort aiming at introducing an automatic tram, that is to say SmartTram. In particular, we concentrate on how the changing role of human (as driver, passenger and member of other user groups) is acknowledged in the design of a new automatic tram. For this reason, we present a human fac tors engineering program for automated trams . The special focus is on how the relevant user groups may be involved in design within the defined program. This approach can be utilized also in other sectors when increasing automation.

Tsuyoshi Miyashita, Ryota Imai, Masaki Kondo, Tadasuke Furuya

Now, in Japan, the coastal shipping industry has problems reducing the seafarers and aging. The closed sea areas in central bays and ports in Japan are navigated by many ships. In these waters, an inexperienced ship operator may make a mistake in judgment due to extreme tension, which may cause a marine accident. The authors made a prototype to provide ship operators on board with maneuvering assistance from shore to solve the problems. We developed a prototype by using wireless and mobile communication, VPN, and web browser. We conducted an actual ship experiment in the natural sea as a verification, discussing its effectiveness. It was shown that maneuvering about the same level as onboard as poss ible. As a result, it was confirmed that this method could be effective by future improvements.

Evan Bryer, Theppatorn Rhujittawiwat, John R. Rose, Colin F. Wilder

The goal of this paper is to modify an existing clustering algorithm with the use of the Hunspell spell checker to specialize it for the use of cleaning early modern European book title data. Duplicate and corrupted data is a constant concern for data analysis, and clustering has been identified to be a robust tool for normalizing and cleaning data such as ours. In particular, our data comprises over 5 million books published in European languages between 150 0 and 1800 in the Machine -Readable Cataloging (MARC) data format from 17,983 libraries in 123 countries. However, as each library individually catalogued their records, many duplicative and inaccurate records exist in the data set. Additionally, each language evolved over the 300-year period we are studying, and as such many of the words had their spellings altered. Without cleaning and normalizing this data, it would be difficult to find coherent trends, as much of the data may be missed in the query. In p revious research, we have identified the use of Prediction by Partial Matching to provide the most increase in base accuracy when applied to dirty data of similar construct to our data set. However, there are many cases in which the correct book title may not be the most common, either when only two values exist in a cluster, or the dirty title exists in more records. In these cases, a language agnostic clustering algorithm would normalize the incorrect title and lower the overall accuracy of the data set. By implementing the Hunspell spell checker into the clustering algorithm, using it to rank clusters by the number of words not found in their dictionary, we can drastically lower the cases of this occurring. Indeed, this ranking algorithm proved to increas e the overall accuracy of the clustered data by as much as 25% over the unmodified Prediction by Partial Matching algorithm.

Pedro Isaías

Editorial preface for Volume 16, Issue 2 of the IADIS International Journal on Computer Science and Information Systems (IJCSIS). This issue brings together peer-reviewed original research contributions covering the wide spectrum of Information Systems, Artificial Intelligence, Software Engineering, and Digital Transformation.