SRT3D: A Sparse Region-Based 3D Object Tracking Approach for the Real World

  title={SRT3D: A Sparse Region-Based 3D Object Tracking Approach for the Real World},
  author={Manuel Stoiber and Martin Pfanne and Klaus H. Strobl and Rudolph Triebel and Alin Albu-Schaffer},
  journal={International Journal of Computer Vision},
  pages={1008 - 1030}
Region-based methods have become increasingly popular for model-based, monocular 3D tracking of texture-less objects in cluttered scenes. However, while they achieve state-of-the-art results, most methods are computationally expensive, requiring significant resources to run in real-time. In the following, we build on our previous work and develop SRT3D, a sparse region-based approach to 3D object tracking that bridges this gap in efficiency. Our method considers image information sparsely along… 

Iterative Corresponding Geometry: Fusing Region and Depth for Highly Efficient 3D Tracking of Textureless Objects

ICG, a novel probabilistic tracker that fuses region and depth information and only requires the object geometry, is proposed, which outperforms the current state of the art with respect to accuracy and robustness.

Large-displacement 3D Object Tracking with Hybrid Non-local Optimization

A hybrid approach combining non- local and local optimizations for different parameters, resulting in efficient non-local search in the 6D pose space is proposed, and a precomputed robust contour-based tracking method is proposed for the pose optimization.

A Multi-body Tracking Framework - From Rigid Objects to Kinematic Structures

The proposed framework considers both tree-like and closed kinematic structures and allows a flexible configuration of joints and constraints and extends ICG, which is a state-of-the-art rigid object tracking algorithm, to multi-body tracking.

Visual Pose Measurement for Automatic Landing on an Aircraft Carrier

A novel airborne non-cooperative method to measure the relative pose for automatic landing using the PP-PicoDet and a texture randomization method to generate high quality data for model training.



3D object tracking via boundary constrained region-based model

A novel energy functional is defined which considers the discrimination between statistical appearance models and the coherence among neighboring pixels simultaneously and a particle filter-like stochastic optimization technique is adopted to solve the energy functional.

Real-Time Model-Based Rigid Object Pose Estimation and Tracking Combining Dense and Sparse Visual Cues

A novel model-based method for estimating and tracking the six-degrees-of-freedom (6DOF) pose of rigid objects of arbitrary shapes in real-time is proposed, both highly accurate and robust to noise and occlusions.

2D-3D Pose Estimation of Heterogeneous Objects Using a Region Based Approach

Inspired by ideas from local active contours, a framework for simultaneous image segmentation and pose estimation using multiple local appearance models is proposed, capable of capturing spatial variation in statistical properties, where global appearance models are limited.

A Sparse Gaussian Approach to Region-Based 6DoF Object Tracking

This work proposes a novel, highly efficient sparse approach to regionbased 6DoF object tracking that requires only a monocular RGB camera and the 3D object model and derives a novel mathematical proof that shows that the proposed likelihood function follows a Gaussian distribution.

Combining 3D Model Contour Energy and Keypoints for Object Tracking

A new combined approach for monocular model-based 3D tracking that eliminates numerous issues of keypoint-based and edge-based approaches and proposes a technique and search area constraints that allow overcoming the local optima and taking into account information obtained through key point-based pose estimation.

Robust 3D Pose Estimation and Efficient 2D Region-Based Segmentation from a 3D Shape Prior

A novelty of the proposed contribution is to use a unique 3D model surface of the object, instead of learning a large collection of 2D shapes to accommodate for the diverse aspects that a 3D object can take when imaged by a camera.

6-DOF Model Based Tracking via Object Coordinate Regression

This work builds on a recently developed state-of-the-art system for single image 6D pose estimation of known 3D objects, using the concept of so-called 3D object coordinates, and creates a new dataset, which will be made publicly available.

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

This paper designs a new strategy to define local regions, which is simple yet efficient in constructing discriminative local color histograms, and proposes to define multiple overlapped, fan-shaped regions according to polar coordinates.

3D Object Tracking with Adaptively Weighted Local Bundles

3D object tracking with monocular RGB images faces many challenges in real environments, and it is observed that the color and edge features are complementary for different cases, and thus a fuse to fuse them for improving tracking robustness is proposed.

Region-based pose tracking with occlusions using 3D models

A framework in which the pose of 3-D models is found by minimising the 2-D projection error through minimisation of an energy function depending on the pose parameters is proposed, which makes it possible to handle occlusions and self-occlusions by tracking multiple objects and object parts simultaneously.