In today’s digital world, we are committed to digitizing thousands of handwritten transcriptions to preserve their content. Historical Arabic Handwritten Text Recognition (HAHTR) remains a challenge for computer vision systems, due to the many difficulties inherently associated with document image quality and the complexity of Arabic script. In this work, we address the problem of recognizing historical Arabic documents that adapts to different writing styles and degrees of legibility. We developed a system that is able to recognize a whole page of a historical Arabic handwritten text in two consecutive steps comprising text line detection and recognition he proposed approach performs detection using bounding boxes followed by a neural network-based model for character-level text recognition. However, the lack of data hinders the mass digitization of Arabic historical documents. Therefore, we provide a new and freely available dataset, focusing on diverse handwriting styles to facilitate a strong generalization of the trained model. This dataset will significantly benefits researchers and practitioners by accelerating progress in the field of HAHTR. Extensive experimental work demonstrates that the recognition models are effective when trained with different sources of data, and having different writing styles does not penalize the model’s ability to generalize but rather enhances it. Additionally, we define and develop a new metric to evaluate model robustness against character misclassification, particularly for characters with similar patterns. The experiments conducted demonstrated that the proposed HAHTR pipeline is accurate and highly generalizable, as well as the validity of bounding box methods for detecting text lines. The training approach with different data sources enabled us to surpass the state-of-the-art results with 5.7% of Character Error Rate (CER) on the KHATT database.

Word embedding is used to represent words for text analysis. It plays an essential role in many Natural Language Processing (NLP) studies and has hugely contributed to the extraordinary developments in the field in the last few years. In Arabic, diacritic marks are a vital feature for the readability and understandability of the language. Current Arabic word embeddings are non-diacritized. In this article, we aim to develop and compare word embedding models based on diacritized and non-diacritized corpora to study the impact of Arabic diacritization on word embeddings. We propose evaluating the models in four different ways: clustering of the nearest words; morphological semantic analysis; part-of-speech tagging; and semantic analysis. For a better evaluation, we took the challenge to create three new datasets from scratch for the three downstream tasks. We conducted the downstream tasks with eight machine learning algorithms and two deep learning algorithms. Experimental results show that the diacritized model exhibits a better ability to capture syntactic and semantic relations and in clustering words of similar categories. Overall, the diacritized model outperforms the non-diacritized model. We obtained some more interesting findings. For example, from the morphological semantics analysis, we found that with the increase in the number of target words, the advantages of the diacritized model are also more obvious, and the diacritic marks have more significance in POS tagging than in other tasks.

Voice over IP (VoIP) is finding its way into several applications, but its security concerns still remain. This paper shows how a new self-adaptive steganographic method can ensure the security of covert VoIP communications over the Internet. In this study an Active Voice Period Detection algorithm is devised for PCM codec to detect whether a VoIP packet carries active or inactive voice data, and the data embedding location in a VoIP stream is chosen randomly according to random sequences generated from a logistic chaotic map. The initial parameters of the chaotic map and the selection of where to embed the message are negotiated between the communicating parties. Steganography experiments on active and inactive voice periods were carried out using a VoIP communications system. Performance evaluation and security analysis indicates that the proposed VoIP steganographic scheme can withstand statistical detection, and achieve secure real-time covert communications with high speech quality and negligible signal distortion.

Reversible and imperceptible watermarking is recognized as a robust approach to confirm the integrity and authenticity of medical images and to verify that alterations can be detected and tracked back. In this paper, a novel blind reversible watermarking approach is presented to detect intentional and unintentional changes within brain Magnetic Resonance (MR) images. The scheme segments images into two parts; the Region of Interest (ROI) and the Region of Non Interest (RONI). Watermark data is encoded into the ROI using reversible watermarking based on the Difference Expansion (DE) technique. Experimental results show that the proposed method, whilst fully reversible, can also realize a watermarked image with low degradation for reasonable and controllable embedding capacity. This is fulfilled by concealing the data into ‘smooth’ regions inside the ROI and through the elimination of the large location map required for extracting the watermark and retrieving the original image. Our scheme delivers highly imperceptible watermarked images, at 92.18–99.94 dB Peak Signal to Noise Ratio (PSNR) evaluated through implementing a clinical trial based on relative Visual Grading Analysis (relative VGA). This trial defines the level of modification that can be applied to medical images without perceptual distortion. This compares favorably to outcomes reported under current state-of-art techniques. Integrity and authenticity of medical images are also ensured through detecting subsequent changes enacted on the watermarked images. This enhanced security measure, therefore, enables the detection of image manipulations, by an imperceptible approach, that may establish increased trust in the digital medical workflow.

This paper focuses on the items/users representation in the domain of recommender systems. These systems compute similarities between items (and/or users) to recommend new items to users based on their previous preferences. It is often useful to consider the characteristics (a.k.a features or attributes) of the items and/or users. This represents items/users by vectors that can be very large, sparse and space-consuming. In this paper, we propose a new accurate method for representing items/users with low size data structures that relies on two concepts: (1) item/user representation is based on bloom filter vectors, and (2) the usage of these filters to compute bitwise AND similarities and bitwise XNOR similarities. This work is motivated by three ideas: (1) detailed vector representations are large and sparse, (2) comparing more features of items/users may achieve better accuracy for items similarities, and (3) similarities are not only in common existing aspects, but also in common missing aspects. We have experimented this approach on the publicly available MovieLens dataset. The results show a good performance in comparison with existing approaches such as standard vector representation and Singular Value Decomposition (SVD).

Identifying semantic relations is a crucial step in discourse analysis and is useful for many applications in both language and speech technology. Automatic detection of Causal relations therefore has gained popularity in the literature within different frameworks. The aim of this article is the automatic detection and extraction of Causal relations that are explicitly expressed in Arabic texts. To fulfill this goal, a Pattern Recognizer model was developed to signal the presence of cause--effect information within sentences from nonspecific domain texts. This model incorporates approximately 700 linguistic patterns so that parts of the sentence representing the cause and those representing the effect can be distinguished. The patterns were constructed based on different sets of syntactic features by analyzing a large untagged Arabic corpus. In addition, the model was boosted with three independent algorithms to deal with certain types of grammatical particles that indicate causation. With this approach, the proposed model achieved an overall recall of 81% and a precision of 78%. Evaluation results revealed that the justification particles play a key role in detecting Causal relations. To the best of our knowledge, no previous studies have been dedicated to dealing with this type of relation in the Arabic language.

The treatment of complex questions with explanatory answers involves searching for arguments in texts. Because of the prominent role that discourse relations play in reflecting text producers’ intentions, capturing the underlying structure of text constitutes a good instructor in this issue. From our extensive review, a system for automatic discourse analysis that creates full rhetorical structures in large-scale Arabic texts is currently unavailable. This is due to the high computational complexity involved in processing a large number of hypothesized relations associated with large texts. Therefore, more practical approaches should be investigated. This article presents a new Arabic Text Parser oriented for question-answering systems dealing with لماذا “why” and كيف “how to” questions. The Text Parser presented here considers the sentence as the basic unit of text and incorporates a set of heuristics to avoid computational explosion. With this approach, the developed question-answering system reached a significant improvement over the baseline with a Recall of 68% and MRR of 0.62.

The investigation reported in this paper is part of a large study comparing the online behaviour of Saudis living in Saudi Arabia and those living in the UK. The study attempts to answer a more general question on whether the environment plays an important role and affects the behavior of online shoppers. The environment here has a wider coverage and includes both technological and social aspects. This paper deals only with the research conducted in the UK and attempts to understand the behaviour and perception of Business to Customer E-Commerce among Saudis living in the UK. This study attempts to evaluate the impact of the environmental change on the online shopping behaviour. Quantitative data was collected from 169 Saudi participants who live in the UK. The key factors tested were: Internet usage, satisfaction on using the Internet, security and payment, integrity, and affiliation and reputation of online merchants. The findings show that there is a high number of Saudi citizens living in the UK that are engaged with E-Commerce and are consistently buying online. These early findings suggest that the environment plays a role in changing the shopping habits of online shoppers.

Early phases of software development are known to be problematic, difficult to manage and errors occurring during these phases are expensive to correct. Many systems have been developed to aid the transition from informal Natural Language requirements to semi-structured or formal specifications. Furthermore, consistency checking is seen by many software engineers as the solution to reduce the number of errors occurring during the software development life cycle and allow early verification and validation of software systems. However, this is confined to the models developed during analysis and design and fails to include the early Natural Language requirements. This excludes proper user involvement and creates a gap between the original requirements and the updated and modified models and implementations of the system. To improve this process, we propose a system that generates Natural Language specifications from UML class diagrams. We first investigate the variation of the input language used in naming the components of a class diagram based on the study of a large number of examples from the literature and then develop rules for removing ambiguities in the subset of Natural Language used within UML. We use WordNet, a linguistic ontology, to disambiguate the lexical structures of the UML string names and generate semantically sound sentences. Our system is developed in Java and is tested on an independent though academic case study.

Recent years have seen much debate about the appropriate content of software engineering (SE) programs and how they relate to computer science (CS) programs, culminating in the distinguishing knowledge areas identified in the ACM/IEEE CS and SE curricula. Given these publications, a reasonable question to ask is: how do current SE programs differ from CS programs and to what extent do the differences reflect the characterizing features given in the ACM/IEEE curricula? We aim to answer these questions for SE programs offered in England. The content of a third of the SE programs in England are analyzed and summarized with respect to the knowledge areas of both the ACM/IEEE CS and SE curricula. The results reveal interesting features; such as intelligent systems is a more distinguishing feature between the CS and SE programs than the expected knowledge areas given in the SE curriculum. The main finding is that there are relatively few differences between existing SE and CS programs offered in England. We conclude with a discussion of the reasons for this situation and its likely implications.

The advent of the WWW and distributed information systems have made it possible to share documents between different users and organisations. However, this has created many problems related to the security, accessibility, right and most importantly the consistency of documents. It is important that the people involved in the documents management process have access to the most up-to-date version of documents, retrieve the correct documents and should be able to update the documents repository in such a way that his or her document are known to others. In this paper we propose a method for organising, storing and retrieving documents based on similarity contents. The method uses techniques based on information retrieval, document indexation and term extraction and indexing. This methodology is developed for the E-Cognos project which aims at developing tools for the management and sharing of documents in the construction domain.

The amount of information available on the Internet is currently growing at an incredible rate. However, the lack of efficient indexing is still a major barrier to effective information retrieval on the Web. This paper presents the design of an intelligent agent for content-based indexing and retrieval of relevant documents from a large collection such as the Internet. The agent aims at improving the quality of retrieval by capturing the semantics of the documents. It performs the conventional keyword based indexing and introduces a thematic relationship between parts of text using natural language understanding and a linguistics theory called rhetorical structure theory. The agent described in this paper will be implemented and compared against several indexing systems. It is expected to produce a satisfactory improvement over existing techniques.

Formal methods provide an approach in which design steps can be shown to satisfy a specification. However, if a formal specification is wrong, then although the design steps may satisfy the formal specification, they are unlikely to satisfy the requirements of the system. Since most users are unfamiliar with formal methods, requirements specifications are often written in English. Such requirements, expressed in English, are then somehow translated to formal specifications. This transition has some potential for introducing errors and inconsistencies. In this paper we propose an interactive approach to proceeding from an informal specification to a formal specification in a systematic manner. The approach uses research in the area of natural language understanding to analyse English specifications in order to detect ambiguities and to generate an entity relationship model. The entity relationship model is then used as a basis for producing VDM data types and the specifications of some common operations. We illustrate the effectiveness of our approach by applying it to the specification of part of a route planning database system.