Search

Natural disasters, like pandemics and earthquakes, are some of the main causes of distress and casualties. Governmental crisis management processes are crucial when dealing with these types of problems. Social media platforms are among the main sources of information regarding current events and public opinion. So, they have been used extensively to aid disaster detection and prevention efforts. Therefore, there is always a need for better automatic systems that can detect and classify disaster data of social media. In this work, we propose enhanced Arabic disaster data classification models. The suggested models utilize domain adaptation to provide state-of-the-art accuracy. We used a standard dataset of Arabic disaster data collected from Twitter for testing the proposed models. Experimental results show that the provided models significantly outperform the previous state-of-the-art results.

In Saudi Arabia, several mitigating measures were implemented in response to the COVID-19 pandemic, including the creation of COVID-19 mobile applications (apps) for public use. The Saudi government has made the use of these apps mandatory for its citizens and residents. However, it is essential to explore the perception that common users have regarding using these apps in terms of usability and user experience. Therefore, this paper assesses user experience in terms of effectiveness, efficiency, and user satisfaction with the usability of the Saudi COVID-19 apps. The reviews of five mobile apps launched by the Saudi Data and AI Authority (SDAIA) and the Ministry of Health in the Apple Store were extracted using an online tool and analyzed using the content analysis method. The number of collected reviews was 29 for Sehha, 406 for Sehhaty, 442 for Mawid, 107 for Tabaud, and 1338 for Tawakkalna. The results of the study showed that Mawid (82%) and Tabaud (81%) had the highest usability of all the apps studied. Sehha (138%) and Sehhaty (−107%) received the lowest usability scores, followed by Tawakkalna (22%). Based on these results, we identified several usability issues with each app. Some of the main problems reported by users were increased battery drain, lack of privacy, and technical issues.

A mispronunciation of Arabic short vowels can change the meaning of a complete sentence. For this reason, both the students and teachers of Classical Arabic (CA) are required extra practice for correcting students’ pronunciation of Arabic short vowels. That makes the teaching and learning task cumbersome for both parties. An intelligent process of students’ evaluation can make learning and teaching easier for both students and teachers. Given that online learning has become a norm these days, modern learning requires assessment by virtual teachers. In our case, the task is about recognizing the exact pronunciation of Arabic alphabets according to the standards. A major challenge in the recognition of precise pronunciation of Arabic alphabets is the correct identification of a large number of short vowels, which cannot be dealt with using traditional statistical audio processing techniques and machine learning models. Therefore, we developed a model that classifies Arabic short vowels using Deep Neural Networks (DNN). The model is constructed from scratch by: (i) collecting a new audio dataset, (ii) developing a neural network architecture, and (iii) optimizing and fine‐tuning the developed model through several iterations to achieve high classification accuracy. Given a set of unseen audio samples of uttered short vowels, our proposed model has reached the testing accuracy of 95.77%. We can say that our results can be used by the experts and researchers for building better intelligent learning support systems in Arabic speech processing.

Enterprise Resource Planning (ERP) is a frequently used system among organizations to automate their workflows, and companies’ performances are highly dependent on the ERP system. The usability issues of ERP systems may cause performance degradation, resulting in the company’s loss in terms of cost. Previously, several studies reported many usability problems of ERP systems. It can be helpful for the developers and designers of ERP systems to use design recommendations as a quick reference to avoid recurrent usability problems of ERP systems. Currently, this area lacks effective consolidation of the previously reported usability problems data. This paper presents a unique approach to developing a precise checklist of ERP usability problems using the topic modeling technique. Our analysis found six different usability problem-related topics that can be generalized for various ERP systems. We have successfully validated our checklist in three different usability studies of ERP systems. The most found usability problems are “difficulty searching and finding desired item/information in interface and error handling” and “missing data and information”. The outcome of our paper is the provision of recommendations to avoid the usability problems of ERP systems and help organizations efficiently prevent frequent issues during the development and maintenance of ERP systems.

Disasters and crises are inevitable in this world. In the aftermath of a disaster, a society’s overall growth, resources, and economy are greatly affected as they cause damages from minor to huge proportions. Around the world, countries are interested in improving their emergency decision-making. The institutions are paying attention to collecting different types of data related to crisis information from various resources, including social media, to improve their emergency response. Previous efforts have focused on collecting, extracting, and classifying crisis data from text, audio, video, or files; however, the development of user-friendly multimodal disaster data dashboards to support human-to-system interactions during an emergency response has received little attention. Our paper seeks to fill this gap by proposing usable designs of interactive dashboards to present multimodal disaster information. For this purpose, we first investigated social media data and metadata for the required elicitation and analysis purposes. These requirements are then used to develop interactive multimodal dashboards to present complex disaster information in a usable manner. To validate our multimodal dashboard designs, we have conducted a heuristic evaluation. Experts have evaluated the interactive disaster dashboards using a customized set of heuristics. The overall assessment showed positive feedback from the evaluators. The proposed interactive multi-modal dashboards complement the existing techniques of collecting textual, image, audio, and video emergency information and their classifications for usable presentation. The contribution will help the emergency response personnel in terms of useful information and observations for prompt responses to avoid significant damage.

Social media platforms have proven to be effective for information gathering during emergency events caused by natural or human-made disasters. Emergency response authorities, law enforcement agencies, and the public can use this information to gain situational awareness and improve disaster response. In case of emergencies, rapid responses are needed to address victims' requests for help. The research community has developed many social media platforms and used them effectively for emergency response and coordination in the past. However, most of the present deployments of platforms in crisis management are not automated, and their operational success largely depends on experts who analyze the information manually and coordinate with relevant humanitarian agencies or law enforcement authorities to initiate emergency response operations. The seamless integration of automatically identifying types of urgent needs from millions of posts and delivery of relevant information to the appropriate agency for timely response has become essential. This research project aims to develop a generalized Information Technology (IT) solution for emergency response and disaster management by integrating social media data as its core component. In this paper, we focused on text analysis techniques which can help the emergency response authorities to filter through the sheer amount of information gathered automatically for supporting their relief efforts. More specifically, we applied state-of-the-art Natural Language Processing (NLP), Machine Learning (ML), and Deep Learning (DL) techniques ranging from unsupervised to supervised learning for an in-depth analysis of social media data for the purpose of extracting real-time information on a critical event to facilitate emergency response in a crisis. As a proof of concept, a case study on the COVID-19 pandemic on the data collected from Twitter is presented, providing evidence that the scientific and operational goals have been achieve

The recent surge of social media networks has provided a channel to gather and publish vital medical and health information. The focal role of these networks has become more prominent in periods of crisis, such as the recent pandemic of COVID-19. These social networks have been the leading platform for broadcasting health news updates, precaution instructions, and governmental procedures. They also provide an effective means for gathering public opinion and tracking breaking events and stories. To achieve location-based analysis for social media input, the location information of the users must be captured. Most of the time, this information is either missing or hidden. For some languages, such as Arabic, the users’ location can be predicted from their dialects. The Arabic language has many local dialects for most Arab countries. Natural Language Processing (NLP) techniques have provided several approaches for dialect identification. The recent advanced language models using contextual-based word representations in the continuous domain, such as BERT models, have provided significant improvement for many NLP applications. In this work, we present our efforts to use BERT-based models to improve the dialect identification of Arabic text. We show the results of the developed models to recognize the source of the Arabic country, or the Arabic region, from Twitter data. Our results show 3.4% absolute enhancement in dialect identification accuracy on the regional level over the state-of-the-art result. When we excluded the Modern Standard Arabic (MSA) set, which is formal Arabic language, we achieved 3% absolute gain in accuracy between the three major Arabic dialects over the state-of-the-art level. Finally, we applied the developed models on a recently collected resource for COVID-19 Arabic tweets to recognize the source country from the users’ tweets. We achieved a weighted average accuracy of 97.36%, which proposes a tool to be used by policymakers to support country

The pronunciation of the Arabic language is required all articulatory phonetics organs to formulate the correct sounds of a word. It is challenging for non-native Arabic speakers to learn to recite the Holy Quran with correct “Tajweed” rules and pronunciation. The limited contributions are made in the development of classic Arabic short vowels dataset that hinder the development of speech recognition system to facilitate the learner during their Holy Quran learning process. Therefore, it is required to have a collection of audio classic Arabic datasets that can help speech recognition and mispronouncing detection of the classic Arabic speech. In this paper, we aim to collect the classical Arabic alphabet with short vowels. Short vowels are an essential part of the Arabic language. One word of the Arabic language consists of at least one or two short vowels. First, we start with requirement gathering for the classical Arabic short vowels. Our primary focus is to record and process the collected audio dataset. The first release of the audio dataset collected consists of 2892 Arabic alphabet short vowels. A significant effort is applied in preprocessing of the dataset consisting of 84 classes of the Arabic alphabet short vowels. Then, the dataset is tested using a sequential convolution neural network (CNN) on 312 phonemes of the collected Arabic Alphabet /a/ "Alif" with short vowels. The result shows that CNN gives high testing accuracy of 100% and a loss of 0.27.

Tomato production is affected by various threats, including pests, pathogens, and nutritional deficiencies during its growth process. If control is not timely, these threats affect the plant-growth, fruit-yield, or even loss of the entire crop, which is a key danger to farmers’ livelihood and food security. Traditional plant disease diagnosis methods heavily rely on plant pathologists that incur high processing time and huge cost. Rapid and cost-effective methods are essential for timely detection and early intervention of basic food threats to ensure food security and reduce substantial economic loss. Recent developments in Artificial Intelligence (AI) and computer vision allow researchers to develop image-based automatic diagnostic tools to quickly and accurately detect diseases. In this work, we proposed an AI-based approach to detect diseases in tomato plants. Our goal is to develop an end-to-end system to diagnose essential crop problems in real-time, ensuring high accuracy. This paper employs various deep learning models to recognize and predict different diseases caused by pathogens, pests, and nutritional deficiencies. Various Convolutional Neural Networks (CNNs) are trained on a large dataset of leaves and fruits images of tomato plants. We compared the performance of ShallowNet (a shallow network trained from scratch) and the state-of-the-art deep learning network (models are fine-tuned via transfer learning). In our experiments, DenseNet consistently achieved high performance with an accuracy score of 95.31% on the test dataset. The results verify that deep learning models with the least number of parameters, reasonable complexity, and appropriate depth achieve the best performance. All experiments are implemented in Python, utilizing the Keras deep learning library backend with TensorFlow.

Melanoma is a type of skin cancer, one of the fatal diseases that appear as an abnormal growth of skin cells and the lesion part often looks like a mole on the skin. Early detection of melanoma from skin lesion by means of screening is an important step towards a reduction in mortality. For this purpose, numerous automatic melanoma diagnosis models based on image processing and machine learning techniques are available for computer-based applications (CBA) and smartphone-based applications (SBA). Since, the smartphones are available as most accessible and easiest methods with built-in camera option, SBA are preferred over CBA. In this paper, we explored the available literature and highlighted the challenges of SBA in terms of execution time due to the limited computing power of smartphones. To resolve this issue of storage of the smartphones, we proposed to develop an SBA that can seamlessly process the image data on the cloud instead of local hardware of the smartphone. Therefore, we designed a study to build a machine learning model of melanoma diagnosis to measure the time taken in preprocessing, segmentation, feature extraction, and classification on the cloud and compared the results with the processing time on the smartphone's local machine. The results showed there is a significant difference of p value < 0.001 on the average processing time taken on both environments. As the processing on the cloud is more efficient. The findings of the proposed research will be helpful for the developers to decide the processing platform while developing smartphone applications for automatic melanoma diagnosis.

This paper aims to improve warning signals of the existing drowsiness detection systems in terms of efficiency and effectiveness. For this purpose, firstly, we exploit visual, audio and tactile modalities to discover the most suitable signal for alerting the car drivers in drowsiness detection systems. Secondly, the resultant signals are merged into various combinations to form multimodal warnings. From this, we got four different combinations. Furthermore, we have done an experimental evaluation in order to identify the most effective and efficient multimodal warnings for drowsiness detection systems. Results from our observational studies showed that the most suitable unimodal signals are as follows: (1) 40 ms on and 50 ms off the flashlight of smartphone for visual warning, (2) Apple iOS Ringtones-Alarm for the audio warning, and (3) 1 pulse in a second with 500 ms inter-stimulus interval for the vibrotactile warning. From the experimental study, we found that the most efficient and preferable warning for alerting the drowsy driver is multimodal signal consisting of visual, audio, and tactile modalities. This paper is helpful for the car industries to develop the effective warnings of drowsiness detection systems, furthermore, it helps in adopting these systems and in preventing accidents.

This paper aims to improve warning signals of the existing drowsiness detection systems in terms of efficiency and effectiveness. For this purpose, firstly, we exploit visual, audio and tactile modalities to discover the most suitable signal for alerting the car drivers in drowsiness detection systems. Secondly, the resultant signals are merged into various combinations to form multimodal warnings. From this, we got four different combinations. Furthermore, we have done an experimental evaluation in order to identify the most effective and efficient multimodal warnings for drowsiness detection systems. Results from our observational studies showed that the most suitable unimodal signals are as follows: (1) 40 ms on and 50 ms off the flashlight of smartphone for visual warning, (2) Apple iOS Ringtones-Alarm for the audio warning, and (3) 1 pulse in a second with 500 ms inter-stimulus interval for the vibrotactile warning. From the experimental study, we found that the most efficient and preferable warning for alerting the drowsy driver is multimodal signal consisting of visual, audio, and tactile modalities. This paper is helpful for the car industries to develop the effective warnings of drowsiness detection systems, furthermore, it helps in adopting these systems and in preventing accidents.

A tactile interface is an alternative channel of communication, which can be utilized to display navigational instructions in cars under high visual and auditory load conditions. We investigated the use of a tactile belt for turn-by-turn information presentation in cars, which was originally designed for the route guidance of blind and pedestrian users. Important information artifacts in turn-by-turn route guidance are the distance to an upcoming crossing and the direction to follow. The tactile belt was examined in a pilot study for presenting direction information in a car navigation system. The pilot study was used to explore first ideas of tactile encodings for direction presentation with the tactile belt. The study presented in this article compares these designs systematically for their use in tactile route guidance in the car. For this purpose we conducted an experiment with 10 participants on real urban roads to evaluate three different vibrotactile patterns. The results show that the "two vibrators front design"was significantly different than the "two vibrators side design". The two vibrator front encoding was preferred significantly. The performance of the participants on two vibrators front design was significantly different than the two vibrators side design. The Friedman test showed a significant difference in the usability of all three vibrotactile patterns. A significant difference was found in the ease of use of two vibrators front encoding compared to other encodings. On the basis of quantitative and qualitative results, we conclude that the two vibrators front encoding is comparatively is the better design for presenting the direction information. Our findings will be helpful for the car industry in designing the tactile based car navigation systems.

A tactile interface is an alternative channel of communication, which can be utilized to display navigational instructions in cars under high visual and auditory load conditions. We investigated the use of a tactile belt for turn-by-turn information presentation in cars, which was originally designed for the route guidance of blind and pedestrian users. Important information artifacts in turn-by-turn route guidance are the distance to an upcoming crossing and the direction to follow. The tactile belt was examined in a pilot study for presenting direction information in a car navigation system. The pilot study was used to explore first ideas of tactile encodings for direction presentation with the tactile belt. The study presented in this article compares these designs systematically for their use in tactile route guidance in the car. For this purpose we conducted an experiment with 10 participants on real urban roads to evaluate three different vibrotactile patterns. The results show that the "two vibrators front design"was significantly different than the "two vibrators side design". The two vibrator front encoding was preferred significantly. The performance of the participants on two vibrators front design was significantly different than the two vibrators side design. The Friedman test showed a significant difference in the usability of all three vibrotactile patterns. A significant difference was found in the ease of use of two vibrators front encoding compared to other encodings. On the basis of quantitative and qualitative results, we conclude that the two vibrators front encoding is comparatively is the better design for presenting the direction information. Our findings will be helpful for the car industry in designing the tactile based car navigation systems.

A car navigation system that displays turn-by-turn instructions using a vibro-tactile waist belt aims to avoid the cognitive load that most car navigation systems place on drivers.

A car navigation system that displays turn-by-turn instructions using a vibro-tactile waist belt aims to avoid the cognitive load that most car navigation systems place on drivers.

Tactile displays are an actively studied means to convey large amount of spatial information in the car. Their advantage compared to conventional car navigation systems is their ability to free the driver's visual and auditory senses. Previously the tactile displays were integrated into the seat of a car to present multiple direction information to the driver. However, in the commercial cars the seat is used to provide the vibro-tactile warning signals, so driver might not differentiate between navigation and warning information. Furthermore, the amount of information presented with tactile displays can cause significant cognitive workload, performance degradation and distraction to the driver. In this paper, we explore different methods of encoding multiple directions information with a tactile belt in the car. We compare the vibro-tactile presentation of spatial turn-by-turn information with a conventional car navigation system to measure cognitive workload, performance and distraction of the driver. We found that drivers showed better orientation performance on the tactile display than with the conventional car navigation system. At the same time there was no difference in cognitive workload, performance, and distraction. Thus, a tactile interface can be useful to present more information than simple left or right directions in high load driving conditions in which drivers are required to observe the traffic situation with their visual and auditory senses.

Visual and auditory displays successfully complement each other presenting information in car navigation systems. However, they distract the visual and auditory attention of the driver, which is needed in many primary driving tasks, such as maneuvering the car or observing the traffic. Tactile interfaces can form an alternative way to display spatial information. The way of how exactly information should be presented in a vibro-tactile way is explored rarely. In this paper we investigate three different designs of vibro-tactile stimulation to convey distance information to the driver using a tactile waist belt. We explore the tactile parameters intensity, rhythm, duration, and body location for encoding the distance information. We conduct a comparative experiment on a real navigation scenario in an urban environment to evaluate our designs. In our study we discovered that rhythm and duration are suitable parameters to generate tactile stimulation for encoding distance information. In this way the driver perceives countable vibro-tactile pulses, which indicate the distance in turn by turn instructions. The approach is found be simple way of encoding complex navigational information.