We’re pleased to share that our recent journal paper ” Multi-Frame Spatiotemporal Feature and Hierarchical Learning Approach for No-Reference Screen Content Video Quality Assessment,” which has just been issued this month (October 2025), in IEEE Transactions on Multimedia (TMM). This is a paper collaborating with The Hong Kong Polytechnic University:
- Ngai-Wing Kwong, Yui-Lam Chan, Sik-Ho Tsang, Ziyin Huang and Kin-Man Lam, “Multi-Frame Spatiotemporal Feature and Hierarchical Learning Approach for No-Reference Screen Content Video Quality Assessment,” in IEEE Transactions on Multimedia, vol. 27, pp. 7632-7647, 2025, doi: 10.1109/TMM.2025.3599071.
Paper Link: https://ieeexplore.ieee.org/document/11124545/
Paper Abstract
With remote work and online meetings becoming more common, screen content videos (SCVs) are used a lot more. This creates a big need for ways to check and keep their video quality high. Traditional methods often need the original video as a reference, which isn’t always available. Other existing methods try to assess quality without a reference but only look at simple features and miss important visual and motion information.
We developed a new AI-based system that can evaluate the quality of SCVs without needing the original video. A neural network is designed, namely Dual-Channel Spatiotemporal Convolutional Neural Network (DCST-CNN), which looks at both the main video content and edges to capture important details over space and time. To better understand how the video changes over time, we added a Temporal Pyramid Transformer (TPT) module that combines information from short and long time periods.
Together, these two parts help the system learn detailed video features and predict video quality more accurately. Our tests show this method performs better than existing no-reference approaches, making it very useful for real-world situations like online collaboration and screen sharing, where smooth video quality is essential.
We will keep advancing our collaboration with leading research institutions to develop cutting-edge technologies!
The research team members
Hong Kong Chu Hai College
- Dr Harris Sik-Ho Tsang, Assistant Professor in the Department of Computer Science
The Hong Kong Polytechnic University
- Dr Ngai-Wing Kwong, Postdoctoral Fellow in the Department of Electrical and Electronic Engineering
- Dr Yui-Lam Chan, Associate Professor in the Department of Electrical and Electronic Engineering
- Dr Ziyin Huang, PhD in the Department of Electrical and Electronic Engineering
- Prof. Kin-Man Lam, Professor in in the Department of Electrical and Electronic Engineering
Some photos regarding the academic paper
The overall framework of our proposed neural network model.
The network architecture of our proposed TPT module.
Performance comparison of SCVQA models on two SCVQA databases.
