A filter based approach for inbetweening
This addresses the inbetweening challenge for animators, but it is incremental as it builds on existing neural network methods without major breakthroughs.
The paper tackles the problem of generating intermediate frames for line-drawn animations by training a convolutional neural network to produce smooth transitions without explicit line correspondence or topological analysis, achieving partial success on real animation production data.
We present a filter based approach for inbetweening. We train a convolutional neural network to generate intermediate frames. This network aim to generate smooth animation of line drawings. Our method can process scanned images directly. Our method does not need to compute correspondence of lines and topological changes explicitly. We experiment our method with real animation production data. The results show that our method can generate intermediate frames partially.