Flat-Pack Bench: Evaluating Spatio-Temporal Understanding in Large Vision-Language Models through Furniture Assembly

TL;DR

Flat-Pack Bench is a new benchmark designed to evaluate large vision-language models on detailed spatio-temporal reasoning tasks in furniture assembly videos, exposing their current limitations.

Contribution

The paper introduces a novel benchmark for fine-grained spatio-temporal understanding in videos, focusing on furniture assembly, which is not addressed by existing coarse-grained benchmarks.

Findings

State-of-the-art LVLMs perform poorly on fine-grained reasoning tasks.

LVLMs struggle with temporal ordering and localization in assembly videos.

Current models have limited tracking and spatial interaction understanding.

Abstract

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.