Skip to content

matkrasinski/MobDIS

Repository files navigation

MobDIS - Video frame interpolation using efficient and lightweight convolutional networks


MobDIS is a lightweight and relatively efficient model for video frame interpolation that utilizes Inverted Residuals blocks used in MobileNetV2. It relies on Dense Inverse Search (DIS) method, which is a deterministic algorithm used for estimating optical flow field. Arch DogDance

vimeo_ucf snu_easy_medium snu_hard_extreme


Related work:

Kubas Malwina, Sarwas Grzegorz (2021a). FastRIFE: Optimization of Real-Time Intermediate Flow Estimation for Video Frame Interpolation. arXiv: 2105.13482 [cs.CV]. url: https://arxiv.org/abs/2105.13482.

Jin Xin, Wu Longhai, Shen Guotao, Chen Youxin, Chen Jie, Koo Jayoon, Hahm Cheul-hee (2022). Enhanced Bi-directional Motion Estimation for Video Frame Interpolation. arXiv: 2206.08572 [cs.CV]. url: https://arxiv.org/abs/2206.08572.

Jin Xin, Wu Longhai, Chen Jie, Chen Youxin, Koo Jayoon, Hahm Cheul-hee (2023a). A Unified Pyramid Recurrent Network for Video Frame Interpolation. arXiv: 2211.03456 [cs.CV]. url: https://arxiv.org/abs/2211.03456.


About

MobDIS - Video frame interpolation using efficient and lightweight convolutional networks

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages