Search papers, labs, and topics across Lattice.
The paper introduces Flat-Pack Bench, a new benchmark for evaluating fine-grained spatio-temporal understanding in LVLMs using furniture assembly videos. The benchmark probes temporal ordering, state localization, part mating, and tracking via multiple-choice questions with visual prompts. Experiments demonstrate that current LVLMs struggle with these tasks, revealing limitations in temporal reasoning, tracking, and understanding spatial interactions.
LVLMs can ace ImageNet, but still fail at understanding how to put together an IKEA chair.
The emergence of Large Vision-Language Models (LVLMs) has significantly advanced video understanding capabilities. However, existing benchmarks focus predominantly on coarse-grained tasks such as action segmentation, classification, captioning, and retrieval. Furthermore, these benchmarks often rely on entities that can be easily identified verbally, like household objects, animals, human subjects, etc., limiting their applicability to complex, in-the-wild video scenarios. But, many applications such as furniture assembly, cooking, etc., require step-by-step fine-grained spatio-temporal understanding of the video, which is not sufficiently evaluated in current benchmarks. To address this gap, we introduce Flat-Pack Bench, a novel benchmark centered on furniture assembly tasks. Our benchmark evaluates LVLMs on nuanced tasks, including temporal ordering of assembly actions, temporal localization of assembly state, understanding part mating, and tracking, using multiple-choice questions paired with visual prompts highlighting relevant parts as references for fine-grained questions. Our experiments reveal that state-of-the-art LVLMs struggle significantly with fine-grained spatio-temporal reasoning, highlighting their limitations in effectively leveraging temporal information from videos, limited tracking ability, and understanding of spatial interactions like physical contact.
