OpenFastVC: An Open Source Library for Video Coding Fast Algorithm Implementation

Hang Yuan, Wei Gao

Abstract: Despite the remarkable coding gains exhibited by the recently released new-generation video coding standards, their serious computational complexity will pose a significant challenge in coding latency to practical applications. Therefore, the corresponding low-complexity optimizations assume paramount importance. To facilitate the research in this field, the first open source software library for video coding fast algorithm implementation, namely OpenFastVC, is proposed in this paper. Specifically, OpenFastVC offers the outputting and processing of the intermediate coding information, e.g., the CU partitioning results, which is indispensable to fast algorithm design. To facilitate the integration of the designed algorithms, OpenFastVC also provides application programming interfaces (APIs) for direct control over the encoding process. Moreover, the existing typical fast algorithms are further implemented in OpenFastVC, enabling researchers to evaluate the performance of their algorithm effortlessly. The release of this library is highly favorable for the design, implementation, and evaluation of video coding fast algorithms, thereby fostering the widespread adoption of the new coding standards. The open source library for OpenFastVC is available at https://openi.pcl.ac.cn/OpenCompression/OpenFastVC.