mlsanityData to 14 Oct 2026

O-MaMa: Learning Object Mask Matching between Egocentric and Exocentric Views

arXiv 2506.06026 ↗First posted 6 Jun 2025cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Understanding the world from multiple perspectives is essential for intelligent systems operating together, where segmenting common objects across different views remains an open problem. We introduce a new approach that re-defines cross-image segmentation by treating it as a mask matching task. Our method consists of: (1) A Mask-Context Encoder that pools dense DINOv2 semantic features to obtain discriminative object-level representations from FastSAM mask candidates, (2) an EgoleftrightarrowExo Cross-Attention that fuses multi-perspective observations, (3) a Mask Matching contrastive loss that aligns cross-view features in a shared latent space, and (4) a Hard Negative Adjacent Mining strategy to encourage the model to better differentiate between nearby objects. O-MaMa achieves the state of the art in the Ego-Exo4D Correspondences benchmark, obtaining relative gains of +22% and +76% in the Ego2Exo and Exo2Ego IoU against the official challenge baselines, and a +13% and +6% compared with the SOTA with 1% of the training parameters.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Builds on 68 of its 90 references that are in the map. First posted on arXiv 6 Jun 2025.

Builds on

68 of 90 references.

References · 68

  1. 2017Attention Is All You Needcited by 60,361
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  3. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  4. 2023Segment Anythingcited by 8,234
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  6. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  7. 2017Mask R-CNNcited by 7,277
  8. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
  9. 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
  10. 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
  11. 2021Ego4D: Around the World in 3,000 Hours of Egocentric Videocited by 1,556
  12. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,443
38 more
  1. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,320
  2. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 984
  3. 2018Panoptic Segmentationcited by 802
  4. 2018Scaling Egocentric Vision: The EPIC-KITCHENS Datasetcited by 795
  5. 2023LISA: Reasoning Segmentation via Large Language Modelcited by 776
  6. 2023LightGlue: Local Feature Matching at Light Speedcited by 590
  7. 2022Open-Vocabulary Semantic Segmentation with Mask-adapted CLIPcited by 565
  8. 2023Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectivescited by 490
  9. 2019YOLACT: Real-time Instance Segmentationcited by 443
  10. 2022XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Modelcited by 439
  11. 2019Video Object Segmentation using Space-Time Memory Networkscited by 438
  12. 2023Segment Anything in High Qualitycited by 376
  13. 2019Panoptic-DeepLabcited by 363
  14. 2023Scaling Open-Vocabulary Object Detectioncited by 343
  15. 2024VideoMamba: State Space Model for Efficient Video Understandingcited by 328
  16. 2020DISK: Learning local features with policy gradientcited by 312
  17. 2016LIFT: Learned Invariant Feature Transformcited by 306
  18. 2023RoMa: Robust Dense Feature Matchingcited by 302
  19. 2023A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondencecited by 301
  20. 2022Generalized Decoding for Pixel, Image, and Languagecited by 294
  21. 2023Fast Segment Anythingcited by 290
  22. 2018YouTube-VOS: Sequence-to-Sequence Video Object Segmentationcited by 232
  23. 2022CMX: Cross-Modal Fusion for RGB-X Semantic Segmentation with Transformerscited by 231
  24. 2023EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anythingcited by 188
  25. 2021COTR: Correspondence Transformer for Matching Across Imagescited by 183
  26. 2020Video Panoptic Segmentationcited by 142
  27. 2023ALIKED: A Lighter Keypoint and Descriptor Extraction Network via Deformable Transformationcited by 121
  28. 2020Rolling-Unrolling LSTMs for Action Anticipation from First-Person Videocited by 120
  29. 2020EGO-TOPO: Environment Affordances from Egocentric Videocited by 111
  30. 2020EfficientPS: Efficient Panoptic Segmentationcited by 111
  31. 2019Deep Learning in Video Multi-Object Tracking: A Surveycited by 107
  32. 2024PSALM: Pixelwise SegmentAtion with Large Multi-Modal Modelcited by 101
  33. 2021Ego-Exo: Transferring Visual Representations from Third-person to First-person Videoscited by 99
  34. 2019The 2019 DAVIS Challenge on VOS: Unsupervised Multi-Object Segmentationcited by 97
  35. 2022My View is the Best View: Procedure Learning from Egocentric Videoscited by 81
  36. 2021CATs: Cost Aggregation Transformers for Visual Correspondencecited by 77
  37. 2024OMG-Seg: Is One Model Good Enough For All Segmentation?cited by 75
  38. 2017Learning Video Object Segmentation with Visual Memorycited by 66

50 most cited of 68 All 68 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 4.

7 papers.

  1. 2025Exo2EgoSyn: Unlocking Foundation Video Generation Models for Exocentric-to-Egocentric Video Synthesiscited by 15
  2. 2025V2-SAM: Marrying SAM2 with Multi-Prompt Experts for Cross-View Object Correspondencecited by 14
  3. 2025EgoVITA: Learning to Plan and Verify for Egocentric Video Reasoningcited by 4
  4. 2026Learning Cross-View Object Correspondence via Cycle-Consistent Mask Predictioncited by 3
  5. 2026From Synchrony to Sequence: Exo-to-Ego Generation via Interpolationcited by 1
  6. 2026Revealing Multi-View Hallucination in Large Vision-Language Modelscited by 0
  7. 2026Learning Cross-View Semantic Priors for Single-Reference Unseen Object Pose Estimationcited by 0