Quality of Experience eva‎luation of Multimedia Services by Machine Learning Approach

1/1/1900 12:00:00 AM

Multimedia services are very important in today’s mobile networks and we expect them to grow significantly. With the increase in the generation of mobile networks, consumers’ expectations of multimedia services have increased, which has made the optimization of networks and services a necessity for service providers. In this regard, Quality of Service (QoS) and Quality of Experience (QoE) criteria are very important for mobile network operators. QoE is the subjective satisfaction or overall perception of users when interacting with a particular product or service. Its eva‎luation is done in the form of user feedback through surveys, interviews or other subjective eva‎luation methods. By understanding users’ experiences, service providers can make informed decisions to improve the overall quality and satisfaction of their products. As a result, QoE-based network measurement, monitoring and optimization for video and image is challenging, especially in mobile networks and smartphone users. In this thesis, QoE eva‎luation on the set of images is discussed along with user feedback, and its goal will be to approach the prediction of image quality to the actual quality. In this way, we use deep learning and distributed Federated Learning method (FL) for our eva‎luation and also use the FL model along with our convolutional neural network. Our results show that by pooling models and updating the global model in federated learning and through iterative steps, the accuracy of image quality is gradually improved, and the privacy-preserving nature of federated learning can address users’ concerns about data security and privacy. QoE has a positive effect. By keeping user data on their devices and only sharing model features and updating it, users will feel more confident and comfortable using apps, leading to a better overall experience.

. كيفيت تجربه، كيفيت خدمات، ارزيابي كيفيت تصوير و ويديو، يادگيري عميق، يادگيري فدرالي

. Quality of Experience, Quality of Service, Image and Video Quality Assessment, Deep Learning, Federated Learning

behnaz kabiri

Dr.Abolfazl Diyanat