PinpointQA: A Dataset and Benchmark for Small Object-Centric Spatial Understanding in Indoor Videos
2026-04-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors created PinpointQA, a new dataset to test how well multimodal large language models (MLLMs) can find and describe small objects in indoor videos. It includes over 1,000 scenes and 10,000 question-answer pairs that get harder in four tasks, from checking if an object is there to precisely describing its location. Their tests show current models struggle, especially with detailed spatial predictions. Training models on PinpointQA improved their performance, showing the dataset is useful for both evaluation and learning.
multimodal large language modelsspatial understandingindoor videosobject localizationScanNet datasetquestion answeringspatial reasoningfine-grained spatial descriptionmachine learning benchmark
Authors
Zhiyu Zhou, Peilin Liu, Ruoxuan Zhang, Luyang Zhang, Cheng Zhang, Hongxia Xie, Wen-Huang Cheng
Abstract
Small object-centric spatial understanding in indoor videos remains a significant challenge for multimodal large language models (MLLMs), despite its practical value for object search and assistive applications. Although existing benchmarks have advanced video spatial intelligence, embodied reasoning, and diagnostic perception, no existing benchmark directly evaluates whether a model can localize a target object in video and express its position with sufficient precision for downstream use. In this work, we introduce PinpointQA, the first dataset and benchmark for small object-centric spatial understanding in indoor videos. Built from ScanNet++ and ScanNet200, PinpointQA comprises 1,024 scenes and 10,094 QA pairs organized into four progressively challenging tasks: Target Presence Verification (TPV), Nearest Reference Identification (NRI), Fine-Grained Spatial Description (FSD), and Structured Spatial Prediction (SSP). The dataset is built from intermediate spatial representations, with QA pairs generated automatically and further refined through quality control. Experiments on representative MLLMs reveal a consistent capability gap along the progressive chain, with SSP remaining particularly difficult. Supervised fine-tuning on PinpointQA yields substantial gains, especially on the harder tasks, demonstrating that PinpointQA serves as both a diagnostic benchmark and an effective training dataset. The dataset and project page are available at https://rainchowz.github.io/PinpointQA.