Circular shape detection and tracking in robotic systems includes a wide, promising array of research that aims to develop robotic skills for interacting with realistic environments in chasing circular objects in real-time. To this end, in this work, the authors contribute a new circular object tracking approach using dynamic feature fusion. The tracker estimates the target positions through frame sequences, and it is built on a robust fusion of possible potential target positions in different feature spaces. The proposed tracking strategy includes three main steps to perform the object chasing. First, the features are extracted from input frames. Then, in each feature space, the estimator will find the predicted position of a target. Finally, the dynamic fusing of information from different feature spaces will validate the target position. Experimental results achieved using the sequence of images obtained immediately from a pan-tilt-zoom (PTZ) camera in a real-time object detection and tracking system are provided to validate the proposed approach.
CITATION STYLE
Tran, T. T., & Ha, C. K. (2018). Dynamic Fusion of Color and Shape for Accurate Circular Object Tracking. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10956 LNAI, pp. 496–507). Springer Verlag. https://doi.org/10.1007/978-3-319-95957-3_52
Mendeley helps you to discover research relevant for your work.