From Model Uncertainty to Human Attention: Localization-Aware Visual Cues for Scalable Annotation Review
For practitioners of large-scale data annotation, this work provides a practical way to improve human-in-the-loop labeling efficiency and accuracy by highlighting spatial errors.
The paper introduces a method to visualize spatial uncertainty in AI-assisted annotation, improving label quality and speed in a study with 120 participants. Uncertainty cues redirect annotator effort toward high-uncertainty predictions, achieving higher quality while being faster overall.
High-quality labeled data is essential for training robust machine learning models, yet obtaining annotations at scale remains expensive. AI-assisted annotation has therefore become standard in large-scale labeling workflows. However, in tasks where model predictions carry two independent components, a class label and spatial boundaries, a model may classify an object with high confidence while mislocalizing it. Existing AI-assisted workflows offer annotators no signal about where spatial errors are most likely. Without such guidance, humans may systematically underinspect subtly misplaced boxes. We address this by studying the effect of visualizing spatial uncertainty via a purpose-built interface. In a controlled study with 120 participants, those receiving uncertainty cues achieve higher label quality while being faster overall. A box-level analysis confirms that the cues redirect annotator effort toward high-uncertainty predictions and away from well-localized boxes. These findings establish localization uncertainty as a lever to improve human-in-the-loop annotation. Code is available at https://mos-ks.github.io/MUHA/.