Video Compression using AI and Signal Processing
Description
Digital image and video compression standards played an important role in the rise of digital communications and entertainment technologies and now are even of more importance in addressing the increased bandwidth and storage challenges in broadcasting and streaming industries and the emerging worlds of social media, Virtual Reality (VR) and metaverse. Recently, deep learning gained significant attention for improving video and image compression. To this effect, deep learning techniques have been applied to various stages of the image and video coding pipeline, including intra prediction, inter prediction, residual coding, and entropy coding, among others. The need for improved visual quality in digital media remains a priority. While the latest standards, such as VVC, have made significant advancements in performance, the advancements in technology and the increasing consumer demand for high-quality media content, has led to higher visual quality expectations for images and videos. Hence, the end-to-end deep image and video compression scheme is motivated by the desire to achieve high compression ratios while preserving image and video quality.
Researchers
Selected Publications
-
Enhancing Image Quality by Reducing Compression Artifacts Using Dynamic Window Swin Transformer. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2024.