mlsanityData to 7 Oct 2026

MoEIoU: Rethinking Bounding-Box Regression as a Mixture of Experts

arXiv 2606.00844 ↗First posted 30 May 2026cs.CV cs.AI cs.LGcited by 0

On the map: Mixture-of-experts → Vision mixture-of-experts

Abstract (arXiv)

Bounding-box regression is a fundamental component of object detection, playing a critical role in precise object localization. Existing Intersection-over-Union (IoU)-based loss functions extend the IoU objective by incorporating geometric penalties, such as center-distance and aspect-ratio mismatch, to improve bounding-box regression. However, these penalties typically remain fixed throughout training and do not account for the optimization dynamics in which predicted boxes initially exhibit large center-distance and shape errors, with later stages focusing on improving overlap with the ground truth. To address this limitation, we introduce MoEIoU, a mixture-of-experts based regression loss that jointly models overlap, center alignment, and aspect-ratio mismatch. MoEIoU aggregates these components using a log-sum-exp function, which emphasizes the dominant localization error while maintaining smooth contributions from other terms. Additionally, a curriculum-based weighting schedule is employed to prioritize correcting box position and shape in early training stages and improving overlap in later stages. We evaluated proposed MoEIoU on PASCAL VOC, HRIPCB, and MS COCO using multiple YOLO architectures, along with large-scale simulation experiments. It consistently outperforms standard and recent state-of-the-art losses, demonstrating faster convergence and improved localization accuracy. We further show that this adaptive aggregation improves existing IoU-based losses, yielding consistent gains and providing more effective optimization guidance for bounding-box regression in object detection frameworks.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Vision mixture-of-experts, one of 12 directions in Mixture-of-experts. Builds on 14 of its 24 references that are in the map. First posted on arXiv 30 May 2026.

Builds on

14 of 24 references.

References · 14

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  3. 2015You Only Look Once: Unified, Real-Time Object Detectioncited by 4,829
  4. 2015Fast R-CNNcited by 2,930
  5. 2017Accurate, Large Minibatch SGD: Training ImageNet in 1 Hourcited by 1,479
  6. 2019Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regressioncited by 1,361
  7. 2018Object Detection with Deep Learning: A Reviewcited by 524
  8. 2019Distance-IoU Loss: Faster and Better Learning for Bounding Box Regressioncited by 454
  9. 2016UnitBox: An Advanced Object Detection Networkcited by 205
  10. 2020Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentationcited by 98
  11. 2021Focal and Efficient IOU Loss for Accurate Bounding Box Regressioncited by 57
  12. 2022SIoU Loss: More Powerful Learning for Bounding Box Regressioncited by 49
2 more
  1. 2023Wise-IoU: Bounding Box Regression Loss with Dynamic Focusing Mechanismcited by 30
  2. 2020HoughNet: Integrating near and long-range evidence for bottom-up object detectioncited by 20

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026HI-MoE: Hierarchical Instance-Conditioned Mixture-of-Experts for Object Detection
  2. 2024Aligning Object Detector Bounding Boxes with Human Preference
  3. 2026Mixture-of-Experts Models in Vision: Routing, Optimization, and Generalization
  4. 2024Routers in Vision Mixture of Experts: An Empirical Study
  5. 2025Detect Anything via Next Point Prediction
  6. 2023Mobile V-MoEs: Scaling Down Vision Transformers via Sparse Mixture-of-Experts
  7. 2025MACE: Mixture-of-Experts Accelerated Coordinate Encoding for Large-Scale Scene Localization and Rendering
  8. 2021Open Vocabulary Object Detection with Pseudo Bounding-Box Labels