arXiv:2511.00510v1 Announce Type: new 
Abstract: This paper investigates Multi-Object Tracking (MOT) in panoramic imagery, which introduces unique challenges including a 360{\deg} Field of View (FoV), resolution dilution, and severe view-dependent distortions. Conventional MOT methods designed for narrow-FoV pinhole cameras generalize unsatisfactorily under these conditions. To address panoramic distortion, large search space, and identity ambiguity under a 360{\deg} FoV, OmniTrack++ adopts a feedback-driven framework that progressively refines perception with trajectory cues. A DynamicSSM block first stabilizes panoramic features, implicitly alleviating geometric distortion. On top of normalized representations, FlexiTrack Instances use trajectory-informed feedback for flexible localization and reliable short-term association. To ensure long-term robustness, an ExpertTrack Memory consolidates appearance cues via a Mixture-of-Experts design, enabling recovery from fragmented tracks and reducing identity drift. Finally, a Tracklet Management module adaptively switches between end-to-end and tracking-by-detection modes according to scene dynamics, offering a balanced and scalable solution for panoramic MOT. To support rigorous evaluation, we establish the EmboTrack benchmark, a comprehensive dataset for panoramic MOT that includes QuadTrack, captured with a quadruped robot, and BipTrack, collected with a bipedal wheel-legged robot. Together, these datasets span wide-angle environments and diverse motion patterns, providing a challenging testbed for real-world panoramic perception. Extensive experiments on JRDB and EmboTrack demonstrate that OmniTrack++ achieves state-of-the-art performance, yielding substantial HOTA improvements of +25.5% on JRDB and +43.07% on QuadTrack over the original OmniTrack. Datasets and code will be made publicly available at https://github.com/xifen523/OmniTrack.

تُعدّ المقدمة الحديثة لـ OmniTrack++ تقدمًا كبيرًا في تتبع الأجسام المتعددة، خاصة في الصور البانورامية. تتناول هذه الطريقة المبتكرة التحديات الفريدة التي يطرحها مجال رؤية بزاوية 360 درجة، مثل تمييع الدقة والتشوهات المعتمدة على الرؤية، والتي أعاقت الطرق التقليدية. من خلال تحسين قدرات التتبع في البيئات المعقدة، لا يُحسن OmniTrack++ الدقة فحسب، بل يفتح أيضًا آفاقًا جديدة للتطبيقات في المراقبة، والمركبات الذاتية القيادة، والروبوتات، مما يجعله تطورًا ملحوظًا في هذا المجال.

La reciente introducción de OmniTrack++ marca un avance significativo en el seguimiento de múltiples objetos, especialmente en imágenes panorámicas. Este enfoque innovador aborda los desafíos únicos que presenta un campo de visión de 360 grados, como la dilución de la resolución y las distorsiones dependientes de la vista, que han obstaculizado los métodos tradicionales. Al mejorar las capacidades de seguimiento en entornos complejos, OmniTrack++ no solo mejora la precisión, sino que también abre nuevas posibilidades para aplicaciones en vigilancia, vehículos autónomos y robótica, lo que lo convierte en un desarrollo notable en el campo.

La récente introduction d'OmniTrack++ marque une avancée significative dans le suivi multi-objets, en particulier dans les images panoramiques. Cette approche innovante s'attaque aux défis uniques posés par un champ de vision à 360 degrés, tels que la dilution de la résolution et les distorsions dépendantes de la vue, qui ont entravé les méthodes traditionnelles. En améliorant les capacités de suivi dans des environnements complexes, OmniTrack++ non seulement améliore la précision, mais ouvre également de nouvelles possibilités d'applications dans la surveillance, les véhicules autonomes et la robotique, ce qui en fait un développement remarquable dans le domaine.

The recent introduction of OmniTrack++ marks a significant advancement in multi-object tracking, particularly in panoramic imagery. This innovative approach tackles the unique challenges posed by a 360-degree field of view, such as resolution dilution and view-dependent distortions, which have hindered traditional methods. By enhancing tracking capabilities in complex environments, OmniTrack++ not only improves accuracy but also opens up new possibilities for applications in surveillance, autonomous vehicles, and robotics, making it a noteworthy development in the field.

OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback

Was this article worth reading? Share it

LucidQuery AI

DeepTracker

Attentive AI

Open Source Surveillance

Deptho.ai

Eyeye

Ready to build your own newsroom?