Experiments and Evaluation: Benchmarking OW‑VISCap Across Open‑World Video Tasks

This article evaluates OW‑VISCap on open‑ and closed‑world segmentation and dense video object captioning, setting new benchmarks on multiple datasets.