Department of Computer Systems Engineering-Scopes
Permanent URI for this collectionhttps://rda.sliit.lk/handle/123456789/2230
Browse
2 results
Search Results
Publication Embargo Deepfake Audio Detection: A Deep Learning Based Solution for Group Conversations(2020 2nd International Conference on Advancements in Computing (ICAC), SLIIT, 2020-12-10) Wijethunga, R.L.M.A.P.C.; Matheesha, D.M.K.; Al Noman, A.; De Silva, K.H.V.T.A.; Tissera, M.; Rupasinghe, L.The recent advancements in deep learning and other related technologies have led to improvements in various areas such as computer vision, bio-informatics, and speech recognition etc. This research mainly focuses on a problem with synthetic speech and speaker diarization. The developments in audio have resulted in deep learning models capable of replicating naturalsounding voice also known as text-to-speech (TTS) systems. This technology could be manipulated for malicious purposes such as deepfakes, impersonation, or spoofing attacks. We propose a system that has the capability of distinguishing between real and synthetic speech in group conversations.We built Deep Neural Network models and integrated them into a single solution using different datasets, including but not limited to Urban- Sound8K (5.6GB), Conversational (12.2GB), AMI-Corpus (5GB), and FakeOrReal (4GB). Our proposed approach consists of four main components. The speech-denoising component cleans and preprocesses the audio using Multilayer-Perceptron and Convolutional Neural Network architectures, with 93% and 94% accuracies accordingly. The speaker diarization was implemented using two different approaches, Natural Language Processing for text conversion with 93% accuracy and Recurrent Neural Network model for speaker labeling with 80% accuracy and 0.52 Diarization-Error-Rate. The final component distinguishes between real and fake audio using a CNN architecture with 94% accuracy. With these findings, this research will contribute immensely to the domain of speech analysis.Publication Embargo An Integrated Framework for Predicting Health Based on Sensor Data Using Machine Learning(2020 2nd International Conference on Advancements in Computing (ICAC), SLIIT, 2020-12-10) Jayaweera, K.N.; Kallora, K.M.C.; Subasinghe, N.A.C.K.; Rupasinghe, L.; Liyanapathirana, C.According to recent studies, the majority of the world's population shows a lack of concern in their health. As a consequence, the non-communicable disease rate has increased dramatically. Amongst these diseases, heart diseases have caused the most catastrophic situations. Apart from the busy lifestyle, studies also show that stress is another factor that causes these diseases. Therefore, the focus of our research is to provide a user-friendly health monitoring system that causes minimum disturbance to its users. However, many studies have focused on predicting health; very few have focused on its usability. The objective of our research is to predict the possibility of cardiac arrests and the presence of stress in real-time using a wearable device prototype. The system uses biometric signals obtained from the photoplethysmogram sensor embedded in the wearable device to perform real-time predictions. We trained three models using random forest, k-nearest neighbor, and logistic regression classification algorithms to predict sudden cardiac arrests with accuracies 99.93%, 99.10%, and 94.47%, respectively. Further, we trained three additional models to predict stress using the same algorithms with accuracies 99.87%, 96.83%, and 65.00%, respectively. Thus, the results of this study show that an integrated framework, capable of predicting different health-related conditions, through sensor data collected from wearable sensors, is feasible.