# Auditing Quality Filters for Long-Tail Human Data Curation

arXiv: https://arxiv.org/abs/2609.31896 · HTML: https://mlsanity.com/paper/2609.31896/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 26 references that are in the map.
- First posted on arXiv 25 Sep 2026.

## Builds on (most cited first, 19 of 19)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Learning from Synthetic Humans](https://mlsanity.com/paper/1701.01370/) (2017)
- [Scaling Open-Vocabulary Object Detection](https://mlsanity.com/paper/2306.09683/) (2023)
- [Virtual Worlds as Proxy for Multi-Object Tracking Analysis](https://mlsanity.com/paper/1605.06457/) (2016)
- [BEDLAM: A Synthetic Dataset of Bodies Exhibiting Detailed Lifelike Animated Motion](https://mlsanity.com/paper/2306.16940/) (2023)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems](https://mlsanity.com/paper/2011.12945/) (2020)
- [CODA: A Real-World Road Corner Case Dataset for Object Detection in Autonomous Driving](https://mlsanity.com/paper/2203.07724/) (2022)
- [MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?](https://mlsanity.com/paper/2108.09518/) (2021)
- [NeuralAnnot: Neural Annotator for 3D Human Mesh Training Sets](https://mlsanity.com/paper/2011.11232/) (2020)
- [Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation](https://mlsanity.com/paper/2407.07532/) (2024)
- [GeoCalib: Learning Single-image Calibration with Geometric Optimization](https://mlsanity.com/paper/2409.06704/) (2024)
- [3DB: A Framework for Debugging Computer Vision Models](https://mlsanity.com/paper/2106.03805/) (2021)
- [AIDE: An Automatic Data Engine for Object Detection in Autonomous Driving](https://mlsanity.com/paper/2403.17373/) (2024)
- [Who's in and who's out? A case study of multimodal CLIP-filtering in DataComp](https://mlsanity.com/paper/2405.08209/) (2024)
- [The missing link: Developing a safety case for perception components in automated driving](https://mlsanity.com/paper/2108.13294/) (2021)
- [Finding Differences Between Transformers and ConvNets Using Counterfactual Simulation Testing](https://mlsanity.com/paper/2211.16499/) (2022)
- [Comparing Apples and Oranges: Off-Road Pedestrian Detection on the NREC Agricultural Person-Detection Dataset](https://mlsanity.com/paper/1707.07169/) (2017)

## Abstract (arXiv)

Robots on construction sites must detect workers who are kneeling or bending, which we call low poses. These workers can be lost from training datasets during automatic labeling. We study a pipeline that detects people, estimates their body joints using NLF, and groups similar poses. Low poses account for only about 2 percent of the retained examples. This low share may partly reflect the pipeline's quality filter, which rejects examples with low detection confidence or uncertain joint estimates. We examine this filtering using four alternative pose clues: bounding-box shape, vertical body span, pose grouping aligned to the scene's vertical direction, and image appearance. All four suggest that low poses are rejected by the filter more often. Separately, controlled simulated scenes show that a person detector fine-tuned on a public construction dataset misses more workers in these poses even when we correct their bounding box height is matched to that of standing workers. These findings suggest that low poses are scarce and hard to find, and we cannot rely on bounding boxes or poses for long-tail human data curation.