Frame rate up Conversion

Frame rate up conversion
  Frame rate up conversion (FRUC) is a technique that generating a higher frame rate video from a lower frame rate video by inserting interpolated frames into lower frame rate video. One of significant applications is reducing motion blur caused by hold-type display characteristics (i.e. liquid crystal display).


.

Video frame interpolation using deep cascaded network structure
  Recently, many variant methods of Convolutional neural networks have been proposed for video frame interpolation. Most of these methods tried to predict dense motion flows or kernels for moving objects. In this study, we exhaustively analyze the advantages of both motion estimation schemes and propose a cascaded system to maximize the advantages of both the schemes.



[Y. Yang, B. T. Oh, “Video frame interpolation using deep cascaded network structure,” Signal Processing: Image Communication, Nov. 2020.]

Multi-layer frame rate up-conversion using trilateral filtering for multi-view video
  One of the hardest problems in frame interpolation is estimating motion in occluded region. Unlike mono view video system, occluded region is partially accessible in Multiview system. Motivated by this, we proposed novel motion estimation scheme which brings occluded information from neighboring views and reconstruct occluded region. Using the reconstructed information, motion estimation is performed more accurately.

[Y. Yang, J. H. Hong, and B. T. Oh, “Multi-layer frame rate up-conversion using trilateral filtering for multi-view video,” IEEE Trans. on Circuits and Systems for Video Technology, Feb. 2019.]
[Y. Yang, D. Lee, and B. T. Oh,“Frame rate up conversion for multi-view video,” APSIPA Annual Summit and Conference, 2016.]

AI Media LAB