Improving the Perceptual Quality of 2D Animation Interpolation
This work addresses the labor-intensive process of 2D animation for animators, though it appears incremental as it builds on existing interpolation methods with specific enhancements.
The paper tackled the problem of improving perceptual quality in 2D animation interpolation by proposing a new architecture (SoftsplatLite), a Distance Transform Module, and a metric (RRLD) for automated data collection, resulting in better perceptual performance and establishing LPIPS and chamfer distance as more appropriate quality measures than PSNR and SSIM.
Traditional 2D animation is labor-intensive, often requiring animators to manually draw twelve illustrations per second of movement. While automatic frame interpolation may ease this burden, 2D animation poses additional difficulties compared to photorealistic video. In this work, we address challenges unexplored in previous animation interpolation systems, with a focus on improving perceptual quality. Firstly, we propose SoftsplatLite (SSL), a forward-warping interpolation architecture with fewer trainable parameters and better perceptual performance. Secondly, we design a Distance Transform Module (DTM) that leverages line proximity cues to correct aberrations in difficult solid-color regions. Thirdly, we define a Restricted Relative Linear Discrepancy metric (RRLD) to automate the previously manual training data collection process. Lastly, we explore evaluation of 2D animation generation through a user study, and establish that the LPIPS perceptual metric and chamfer line distance (CD) are more appropriate measures of quality than PSNR and SSIM used in prior art.