OmniLottie: Generating Vector Animations via Parameterized Lottie Tokens

TL;DR

OmniLottie is a new framework that leverages a custom tokenizer and pretrained vision-language models to generate high-quality, semantically aligned vector animations from multi-modal instructions, addressing challenges in learning from raw Lottie JSON files.

Contribution

We introduce a Lottie tokenizer and a large-scale dataset, enabling the use of pretrained models for flexible, high-quality vector animation generation from multi-modal inputs.

Findings

OmniLottie produces vivid, semantically aligned animations.

The tokenizer effectively transforms complex JSON into learnable sequences.

Our dataset supports training and evaluation of vector animation models.

Abstract

OmniLottie is a versatile framework that generates high quality vector animations from multi-modal instructions. For flexible motion and visual content control, we focus on Lottie, a light weight JSON formatting for both shapes and animation behaviors representation. However, the raw Lottie JSON files contain extensive invariant structural metadata and formatting tokens, posing significant challenges for learning vector animation generation. Therefore, we introduce a well designed Lottie tokenizer that transforms JSON files into structured sequences of commands and parameters representing shapes, animation functions and control parameters. Such tokenizer enables us to build OmniLottie upon pretrained vision language models to follow multi-modal interleaved instructions and generate high quality vector animations. To further advance research in vector animation generation, we curate MMLottie-2M, a large scale dataset of professionally designed vector animations paired with textual and visual annotations. With extensive experiments, we validate that OmniLottie can produce vivid and semantically aligned vector animations that adhere closely to multi modal human instructions.