TAO-Amodal: A Benchmark for Tracking Any Object Amodally

arXiv:2312.12433v3 Announce Type: replace-cross
Abstract: Amodal perception, the ability to comprehend complete object structures from partial visibility, is a fundamental skill, even for infants. Its significance extends to applications like autonomous driving, where a clear understanding of heavily occluded objects is essential. However, modern detection and tracking algorithms often overlook this critical capability, perhaps due to the prevalence of \textit{modal} annotations in most benchmarks. To address the scarcity of amodal benchmarks, we introduce TAO-Amodal, featuring 833 diverse categories in thousands …

abstract algorithms applications arxiv autonomous autonomous driving benchmark clear cs.ai cs.cv cs.lg detection driving however modern object objects perception significance tao tracking type understanding visibility