Reproduce “NarrLV: Towards a Comprehensive Narrative-Centric Evaluation for Long Video Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
X. Feng, H. Yu, M. Wu, S. Hu, J. Chen, C. Zhu, J. Wu, X. Chu and 1 more
NarrLV evaluates whether generated videos express requested event sequences. It varies prompts across one to six Temporal Narrative Atoms and three types of change, scoring element fidelity, unit coverage, and coherence between adjacent units. Tested models express fewer requested events as requests grow; among unanimous-annotator cases, coverage accuracy is 0.80, versus 0.55 for StoryEval and 0.27 for VBench-2.0.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.