Abstract:Both region-based methods and direct methods have become popular in recent years for tracking the 6-dof pose of an object from monocular video sequences. Region-based methods estimate the pose of the object by maximizing the discrimination between statistical foreground and background appearance models, while direct methods aim to minimize the photometric error through direct image alignment. In practice, region-based methods only care about the pixels within a narrow band of the object contour due to the level-set-based probabilistic formulation, leaving the foreground pixels beyond the evaluation band unused. On the other hand, direct methods only utilize the raw pixel information of the object, but ignore the statistical properties of foreground and background regions. In this paper, we find it beneficial to combine these two kinds of methods together. We construct a new probabilistic formulation for 3D object tracking by combining statistical constraints from region-based methods and photometric constraints from direct methods. In this way, we take advantage of both statistical property and raw pixel values of the image in a complementary manner. Moreover, in order to achieve better performance when tracking heterogeneous objects in complex scenes, we propose to increase the distinctiveness of foreground and background statistical models by partitioning the global foreground and background regions into a small number of sub-regions around the object contour. We demonstrate the effectiveness of the proposed novel strategies on a newly constructed real-world dataset containing different types of objects with ground-truth poses. Further experiments on several challenging public datasets also show that our method obtains competitive or even superior tracking results compared to previous works. In comparison with the recent state-of-art region-based method, the proposed hybrid method is proved to be more stable under silhouette pose ambiguities with a slightly lower tracking accuracy.

Robust Monocular Object Pose Tracking for Large Pose Shift Using 2D Tracking

Robust and Accurate Monocular Pose Tracking for Large Pose Shift.

Robust Visual Tracking Via CAMShift and Structural Local Sparse Appearance Model

Robust monocular 3D object pose tracking for large visual range variation in robotic manipulation via scale-adaptive region-based method

A Robust Monocular 3D Object Tracking Method Combining Statistical and Photometric Constraints

Robust Monocular Model-Based Pose Tracking of Markerless Rigid Objects

Occlusion-Aware Region-Based 3D Pose Tracking of Objects with Temporally Consistent Polar-Based Local Partitioning

Robust Monocular Pose Tracking of Less-Distinct Objects Based on Contour-Part Model

Robust Visual Tracking Via Multiple Discriminative Models with Object Proposals

Multi-features Guided Robust Visual Tracking.

Beyond Local Search: Tracking Objects Everywhere with Instance-Specific Proposals

Seeing Through the Occluders: Robust Monocular 6-DOF Object Pose Tracking via Model-Guided Video Object Segmentation

3D Object Tracking for Rough Models.

Pose Optimization in Edge Distance Field for Textureless 3D Object Tracking.

Robust Monocular Pose Tracking of Less-Distinct Objects Using Contour Part Model

A Direct 3D Object Tracking Method Based on Dynamic Textured Model Rendering and Extended Dense Feature Fields

Robust Edge-Based 3D Object Tracking with Direction-Based Pose Validation.

Estimating Metric Poses of Dynamic Objects Using Monocular Visual-Inertial Fusion

Tracking as recognition: a stable 3D tracking framework

A Robust Framework for 2D Human Pose Tracking with Spatial and Temporal Constraints

OmniPose6D: Towards Short-Term Object Pose Tracking in Dynamic Scenes from Monocular RGB