DeepQoE: A unified Framework for Learning to Predict Video QoE
Huaizheng Zhang, Han Hu, Guanyu Gao, Yonggang Wen and Kyle Guan*
Nanyang Technological University, Singapore
*Nokia Bell Labs, 791 Holmdel Road, Holmdel, NJ, USA
Abstract
- Motivated by the prowess of deep learning (DL) based techniques in prediction, generalization, and representation learning, we develop a novel framework called DeepQoE to predict video QoE. The end-to-end framework first uses a combination of DL techniques (e.g., word embeddings) to extract generalized features. Next, these features are combined and fed into a neural network for representation learning. Such representations serve as inputs for classification or regression tasks. Evaluating the performance of DeepQoE with two datasets, we show that for the small dataset, the accuracy of all state-of-the-art machine learning algorithm is improved by using the representation derived from DeepQoE. For the large dataset, our DeepQoE framework achieves significant performance improvement in comparison to the best baseline method (90.94\% vs. 82.84\%). Moreover, DeepQoE, also released as an open source tool, provides video QoE research much-needed flexibility in fitting different datasets, extracting generalized features, and learning representations.
Paper download
Slide
code
Pre-trained models