Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The Realtime Face Video Enhancement (RTFVE) model addresses the limitations of low bitrate video conferencing by providing a solution that enhances video quality without requiring specialized hardware. This model integrates seamlessly with existing video decoders and operates in real-time on standard CPUs, making it accessible for widespread use. Experimental results demonstrate significant improvements in perceptual quality compared to compressed video baselines across various low bitrate settings.
RTFVE enhances low bitrate video conferencing quality in real-time on standard CPUs, making high-quality video accessible to all users, regardless of hardware limitations.
There's been a surge in adoption of video conferencing applications for both personal and business use cases. However, the bandwidth limitations faced by many users worldwide may restrict the optimal use of such applications. Although deep learning offers a solution for enhancing low bit rate videos, most models today are either hard to incorporate with modern compression standards or require specialized hardware to run such as significant GPUs making these models impractical. To address these issues, we introduce the Realtime Face Video Enhancement (RTFVE) model which can be easily incorporated with any video decoder and can run in realtime on ordinary CPUs. Experiments show that our model improves perceptual quality over the compressed video baseline at multiple low bitrate settings. The source code will be made available at https://github.com/varun-jois/RTFVE.