Video frame interpolation (VFI) aims to generate predictive frames by motion-warping from bidirectional references. Most examples of VFI utilize spatiotemporal semantic information to realize motion estimation and interpolation. However, due to variable acceleration, irregular movement trajectories, and camera movement in real-world cases, they can not be sufficient to deal with non-linear middle frame estimation. In this paper, we present a reformulation of the VFI as a joint non-linear motion regression (JNMR) strategy to model the complicated inter-frame motions. Specifically, the motion trajectory between the target frame and multiple reference frames is regressed by a temporal concatenation of multi-stage quadratic models. Then, a comprehensive joint distribution is constructed to connect all temporal motions. Moreover, to reserve more contextual details for joint regression, the feature learning network is devised to explore clarified feature expressions with dense skip-connection. Later, a coarse-to-fine synthesis enhancement module is utilized to learn visual dynamics at different resolutions with multi-scale textures. The experimental VFI results show the effectiveness and significant improvement of joint motion regression over the state-of-the-art methods. The code is available at https://github.com/ruhig6/JNMR .

High efficiency video coding (HEVC) has been finalized as the most widely utilized video coding standard, jointly developed by ITU-T, VCEG, and MPEG. In HEVC, the quad-tree structure of the coding unit partition is one of the most substantial modules and provides significant coding gains following huge coding time. In this paper, a rate–distortion-based coding unit partition network (RDNet) is proposed to make partition decisions based on the statistical features. RDNet is composed of a prediction sub-network and a target sub-network, where the prediction sub-network is used to predict the CU partition modes of the intra-prediction and the target sub-network is designed to optimize the network parameters by evaluating the rate–distortion cost, respectively. To balance the prediction accuracy and the rate–distortion loss, a parameter-exchanging strategy is applied to control the parameters’ sharing between two networks. Experimental results prove that our model can reduce the encoding time of HEVC by 55.83~71.72% with an efficient BD-BR of 2.876~3.347%, and the ablation study evaluates the ability of our strategy on balancing the trade-off between coding accuracy and inference speed.

Video frame interpolation (VFI) aims to generate predictive frames by warping learnable motions from the bidirectional historical references. Most existing works utilize spatio-temporal semantic information extractor to realize motion estimation and interpolation modeling. However, they insufficiently consider the real mechanistic rationality of generated middle motions. In this paper, we reformulate VFI as a Joint Non-linear Motion Regression (JNMR) strategy to model the complicated motions of inter-frame. Specifically, the motion trajectory between the target frame and the multiple reference frames is regressed by a temporal concatenation of multi-stage quadratic models. ConvLSTM is adopted to construct this joint distribution of complete motions in temporal dimension. Moreover, the feature learning network is designed to optimize for the joint regression modeling. A coarse-to-fine synthesis enhancement module is also conducted to learn visual dynamics at different resolutions through repetitive regression and interpolation. Experimental results on VFI show that the effectiveness and significant improvement of joint motion regression compared with the state-of-the-art methods. The code is available at https://github.com/ruhig6/JNMR.