Autocomplete Hand-drawn Animations

摘要

Hand-drawn animation is a major art form and communicationmedium, but can be challenging to produce. We present a systemto help people create frame-by-frame animations through manualsketches. We design our interface to be minimalistic: it containsonly a canvas and a few controls. When users draw on the canvas,our system silently analyzes all past sketches and predicts whatmight be drawn in the future across spatial locations and temporalframes. The interface also offers suggestions to beautify existingdrawings. Our system can reduce manual workload and improveoutput quality without compromising natural drawing flow and control:users can accept, ignore, or modify such predictions visualizedon the canvas by simple gestures. Our key idea is to extend thelocal similarity method in [Xing et al. 2014], which handles onlylow-level spatial repetitions such as hatches within a single frame,to a global similarity that can capture high-level structures acrossmultiple frames such as dynamic objects. We evaluate our systemthrough a preliminary user study and confirm that it can enhanceboth users’ objective performance and subjective satisfaction.