InfiniDreamer: Arbitrarily Long Human Motion Generation via Segment Score Distillation
This addresses a domain-specific problem for human motion generation, offering an incremental improvement by enabling longer sequences without requiring long training data.
The paper tackles the problem of generating arbitrarily long human motion sequences, which is limited by the lack of long training data, by proposing InfiniDreamer, a framework that assembles sub-motions and refines them using Segment Score Distillation, achieving coherent and contextually aware results as validated in experiments.
We present InfiniDreamer, a novel framework for arbitrarily long human motion generation. InfiniDreamer addresses the limitations of current motion generation methods, which are typically restricted to short sequences due to the lack of long motion training data. To achieve this, we first generate sub-motions corresponding to each textual description and then assemble them into a coarse, extended sequence using randomly initialized transition segments. We then introduce an optimization-based method called Segment Score Distillation (SSD) to refine the entire long motion sequence. SSD is designed to utilize an existing motion prior, which is trained only on short clips, in a training-free manner. Specifically, SSD iteratively refines overlapping short segments sampled from the coarsely extended long motion sequence, progressively aligning them with the pre-trained motion diffusion prior. This process ensures local coherence within each segment, while the refined transitions between segments maintain global consistency across the entire sequence. Extensive qualitative and quantitative experiments validate the superiority of our framework, showcasing its ability to generate coherent, contextually aware motion sequences of arbitrary length.