Next generation video coding algorithm using machine learning and its hardware implementation

2020 Fiscal Year Final Research Report

• Project/Area Number: 19K24347
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: The University of Tokushima
• Principal Investigator: KATAYAMA Takafumi 徳島大学, 大学院社会産業理工学研究部(理工学域), 助教 (70848522)
• Project Period (FY): 2019-08-30 – 2021-03-31
• Keywords: Image processing / Versatile video coding / Machine learning

Outline of Final Research Achievements

With the spread of edge environments for AI and IoT, moving images will continue to play a major role as a method of information transmission. Recently, codec technology is implemented to various devices around us in order to provide accurate and detailed moving images. In the next-generation coding technology, parallel processing and application to ultra-high resolution are difficult. Therefore, a drastic solution is required especially from the viewpoint of hardware implementation.
By completing this work, it will be possible to design a dedicated IC chip that combines the technologies of the next-generation coding method and artificial intelligence.

Free Research Field

Computer science

Academic Significance and Societal Importance of the Research Achievements

先行研究では、動画像符号化と人工知能を組み合わせた手法は提案されていた。しかし、先行研究で提案された方法は、動画像符号化の効率を向上する半面、人工知能回路がより複雑になることからハードウェアへの実装が困難となっていた。本課題に対して、本研究では人工知能回路の提案と動画像符号化に対する親和性の評価を実施した。この課題を解決することで、社会で利用される電子デバイスの映像処理が、今後さらに効率的に実現できることが期待される。