http://hdl.handle.net/10366/151928 2026-04-21T14:16:59Z http://hdl.handle.net/10366/151990 Gender classification is an important biometric task. It has been widely studied in the literature. Face modality is the most studied aspect of human-gender classification. Moreover, the task has also been investigated in terms of different face components such as irises, ears, and the periocular region. In this paper, we aim to investigate gender classification based on the oral region. In the proposed approach, we adopt a convolutional neural network. For experimentation, we extracted the region of interest using the RetinaFace algorithm from the FFHQ faces dataset. We achieved acceptable results, surpassing those that use the mouth as a modality or facial sub-region in geometric approaches. The obtained results also proclaim the importance of the oral region as a facial part lost in the Covid-19 context when people wear facial mask. We suppose that the adaptation of existing facial data analysis solutions from the whole face is indispensable to keep-up their robustness. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151989 Is early cancer detection using deep learning models reliable? The creation of expert systems based on Deep Learning can become an asset for the achievement of an early detection, offering a preliminary diagnosis or a second opinion, as if it were a second specialist, thus helping to reduce the mortality rate of cancer patients. In this work, we study the differences and impact of various optimizers and hyperparameters in a Convolutional Neural Network model, to then be tested on different datasets. The results of the tests are analyzed and an implementation of a cancer classification model is proposed focusing on the different approaches of the selected Optimizers as the best method for the achievement of optimal results in accurately improving the detection of cancerous cells. Cancer, despite being considered one of the biggest health problems worldwide, continues to be a major problem because its cause remains unknown. Regular medical check-ups are not frequent in countries where access to specialized health services is not affordable or easily accessible, leading to detection in more advanced stages when the symptoms are quite visible. To reduce cases and mortality rates ensuring early detection is paramount. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151987 Sentiment Classification is a key area of natural language processing research that is frequently utilized in several industries. The goal of sentiment analysis is to figure out if a product or service received a negative or positive response. Sentiment analysis is widely utilized in several commercial fields to enhance the quality of services (QoS) for goods or services by gaining a better knowledge of consumer feedback. Deep learning provides cutting-edge achievements in a variety of complex fields. The goal of the study is to propose an improved approach for evaluating and categorising sentiments into different groups. This study proposes a novel hybridised model that combines the benefits of deep learning technologies Dual LSTM (Long Short Term Memory) and CNN (Convolution Neural Network) with the word embedding technique. The performance of three distinct word embedding approaches is compared in order to choose the optimal embedding for the proposed model's implementation. In addition, attention-based BiLSTM is used in a multi-convolutional approach. Standard measures were used to verify the validity of the suggested model's performance. The results show that the proposed model has a significantly enhanced accuracy of 96.56%, which is significantly better than existing models. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151988 With enormous evolution in Microelectronics, Wireless Sensor Networks (WSNs) have played a vital role in every aspect of daily life. Technological advancement has led to new ways of thinking and of developing infrastructure for sensing, monitoring, and computational tasks. The sensor network constitutes multiple sensor nodes for monitoring, tracking, and surveillance of remote objects in the network area. Battery replacement and recharging are almost impossible; therefore, the aim is to develop an efficient routing protocol for the sensor network. The Fuzzy Based Cluster Head Selection (FBCHS) protocol is proposed, which partitions the network into several regions based on node energy levels. The proposed protocol uses an artificial intelligence technique to select the Cluster Head (CH) based on maximum node Residual Energy (RE) and minimum distance. The transmission of data to the Base Station (BS) is accomplished via static clustering and the hybrid routing technique. The simulation results of the FBCHS protocol are com- pared to the SEP protocol and show improvement in the stability period and improved overall performance of the network. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151986 Nowadays, information technology is not only widely used in all walks of life but also fully applied in the marketing and advertisement sector. In particular, Artificial Intelligence (AI) has received growing attention worldwide because of its impact on advertising. However, it remains unclear how social media users react to AI advertisements. The purpose of this study is to examine the behavior of social media users towards AI-based advertisements. This study used a qualitative method, including a semi-structured interview. A total of 23 semi-structured interviews were conducted with social media users aged 18 and over, using a purposive sampling method. The interviews lasted between 27.05–50.39 minutes on average (Mean: 37.48 SD: 6.25) between August and October 2021. We categorized the findings of the current qualitative research into three main process themes: I) reception; II) diving; and III) break-point. While 'reception' covers positive and negative sub-themes, 'diving' includes three themes: comparison, timesaving, and leaping. The final theme, 'break-point', represents the decision-making stage and includes negative or positive opinions. This study provides content producers, social media practitioners, marketing managers, advertising industry, AI researchers, and academics with many insights into AI advertising. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151984 As more people drive vehicles, there is a corresponding increase in the number of deaths and injuries that happen due to road traffic accidents. Thus, various solutions have been proposed to reduce the impact of accidents. One of the most popular solutions is autonomous driving, which involves a series of embedded systems. These embedded systems assist drivers by providing crucial information on the traffic environment or by acting to protect the vehicle occupants in particular situations or to aid driving. Autonomous driving has the capacity to improve transportation services dramatically. Given the successful use of visual technologies and the implementation of driver assistance systems in recent decades, vehicles are prepared to eliminate accidents, congestion, collisions, and pollution. In addition, the IoT is a state-of-the-art invention that will usher in the new age of the Internet by allowing different physical objects to connect without the need for human interaction. The accuracy with which the vehicle's environment is detected from static images or videos, as well as the IoT connections and data management, is critical to the success of autonomous driving. The main aim of this review article is to encapsulate the latest advances in vision strategies and IoT technologies for autonomous driving by analysing numerous publications from well-known databases. 2023-01-24T00:00:00Z http://hdl.handle.net/10366/151985 The recent improvements related to the area of electric locomotive, power electronics, assembly processes and manufacturing of machines have increased the robustness and reliability of induction motors. Regardless of the increased availability, the application of induction motors in many fields alleges the need for operating state supervision and condition monitoring. In other words, fault identification at the initial stage helps make appropriate control decisions, influencing product quality as well as providing safety. Inspired by these demands, this work proposes a regression based modeling for the analysis of performance in induction motors. In this approach, the feature extraction process is combined with classification for efficient fault detection. Deep Belief Network (DBN) stacked with multiple Restricted Boltzmann Machine (RBM) is exploited for the robust diagnosis of faults with the adoption of training process. The influences of harmonics over induction motors are identified and the losses are mitigated. The simulation of the suggested approach and its comparison with traditional approaches are executed. An overall accuracy of 99.5% is obtained which in turn proves the efficiency of DBN in detecting faults. 2023-01-24T00:00:00Z