Video acuity assessment in mobile devices

摘要: The quality of mobile videos is usually quantified through the Quality Experience (QoE), which based on network QoS measurements, user engagement, or post-view subjective scores. Such quantifications are not adequate for real-time evaluation. They cannot provide on-line feedback improvement visual acuity, represents actual viewing experience end user. We present a acuity framework makes fast online computations in device and an accurate estimate video QoE. identify study three main causes that impact videos: spatial distortions, types buffering resolution changes. Each them can be accurately modeled using our framework. use machine learning techniques to build prediction model depicts more than 78% accuracy. experimental implementation iPhone 4 5s show proposed feasible deploy devices. Using data corpus over 2852 clips experiments, we validate