Sep 2026· Italian National Conference on Sensors· Vol 26, pp. 5819· 0 citations· 27 references
Medicine
TL;DR
AdaSAM2 is introduced, a framework that formulates adaptive search-region selection as a sequential decision-making problem that employs an event-driven reinforcement learning policy that selects the cropping scale only during initialization and unreliable tracking states, while reusing the previous configuration during stable tracking.
Abstract
The persistent challenge in visual object tracking, particularly for small and fast-moving targets, lies in the trade-off between effective resolution and contextual information. Fixed search regions cannot adapt to variations in target scale and motion, often resulting in degraded target representation and tracking failures. In this paper, we introduce AdaSAM2, a framework that formulates adaptive search-region selection as a sequential decision-making problem. Unlike conventional per-frame heuristics, our approach employs an event-driven reinforcement learning policy that selects the cropping scale only during initialization and unreliable tracking states, while reusing the previous configuration during stable tracking. To handle target disappearance, we further introduce a lost-aware recovery mechanism that combines progressive search-region enlargement with constrained policy re-selection. Extensive experiments on TSFMO, LaTOT, UAV123, and UAVDT demonstrate consistent improvements over the SAMITE baseline. For example, AdaSAM2 improves Success and Precision on TSFMO from 42.6% and 74.0% to 43.9% and 75.3%, respectively, while improving UAV123 Success and Precision from 69.6% and 92.7% to 71.1% and 94.8%. Moreover, the RL policy is activated on only 0.74% of processed frames on TSFMO, resulting in an average overhead of only 0.0085 ms per frame. These results demonstrate that adaptive input-space optimization can improve tracking accuracy while introducing negligible computational overhead.
Diffusion-based detectors begin inference from noisy boxes, whereas tracking-by-detection pipelines usually
localize each video frame independently. This study examines whether propagated tracker boxes can guide a frozen
DiffusionDet detector without retraining. Confirmed tracks are propagated using their latest observ...
Muhammad Mustapha Miko, De-Quan Li· International Journal of Inn...· 0 citations
Autonomous mobile manipulators operating in dynamic environments face significant challenges due to spatial variability, diverse task requirements, and the complexities of natural language understanding. Traditional systems often lack persistent perception, fail to adapt to open-vocabulary commands, and cannot recall s...
Yuning Cao, Qing-Song Xu· Journal of Physics, Conferen...· 0 citations
A new MOT approach focused on dynamic adaptation and collaborative enhancement consisting of a nonlinear adaptive kalman filter that makes adjustments in object states for detecting anomalous motions and a multi-dimensional feature enhancement network that elaborates on appearance-enhancing exploration accumulations ac...
Yuan-Zhe Ji, Ying Guo, Ze-Yu Liu et al.· Applied intelligence (Boston...· 0 citations
Multi-object tracking (MOT) is dominated by the tracking-by-detection paradigm, whose methods typically rely on a small set of hyperparameters that are conventionally chosen by hand. Tuning them requires repeated expert-guided experimentation, while the procedures used to select reported values are often not systematic...
An occlusion-robust target tracking framework that maintains target identity and trajectory continuity through the integration of YOLOv11n object detection, Kalman Filter motion prediction, and occlusion-aware appearance-based re-identification is proposed.
Mais.M Mohammed, Sharifa Mohammed, Hanan Awadh et al.· 0 citations
Multi-object tracking (MOT) involves maintaining consistent target identities as objects dynamically enter and leave a scene. Deterministic approaches, such as tracking-by-detection with data association, produce reproducible results and are computationally efficient, but they rely heavily on motion models and are sens...
T. Van Nguyen, Rasmus G. K. Christiansen, Dirk Kraft et al.· 0 citations
Related blog posts
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 9, 2026
A machine-learning approach called CheMeleon provides a practical way to develop new drugs or other molecules with much less detailed, expensive experimentation.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.