BOP: Benchmark for 6D Object Pose Estimation

By Tomas Hodan, Frank Michel, Eric Brachmann, Wadim Kehl, A. Buch, Dirk Kraft, Bertram Drost, Joel Vidal, Stephan Ihrke, Xenophon Zabulis and 6 more. European Conference on Computer Vision, 2018

544 citations recorded by the source corpus.

We propose a benchmark for 6D pose estimation of a rigid object from a single RGB-D input image. The training data consists of a texture-mapped 3D object model or images of the object in known 6D poses. The benchmark comprises of: (i) eight datasets in a unified format that cover different practical scenarios, including two new datasets focusing on varying lighting conditions, (ii) an evaluation methodology with a pose-error function that deals with pose ambiguities, (iii) a comprehensive evaluation of 15 diverse recent methods that captures the status quo of the field, and (iv) an online evaluation system that is open for continuous submission of new results. The evaluation shows that methods based on point-pair features currently perform best, outperforming template matching methods, learning-based methods and methods based on 3D local features. The project website is available at bop.felk.cvut.cz.

Subfields: Aerospace Engineering, Biomedical Engineering, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/57e9ef61cf09274f09a4fe83dbdb38e0da4609d2

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Biomedical Engineering

In Control and Systems Engineering

Citations in the corpus

FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (2023, background): “• Average recall (AR) of VSD, MSSD and MSPD metrics introduced in the BOP challenge [25].”
Pix2Pose: Pixel-Wise Coordinate Regression of Objects for 6D Pose Estimation (2019, methodology): “The performance is also better than the best learning-based methods [2, 16] in the benchmark [11].”
BOP Challenge 2020 on 6D Object Localization (2020, methodology): “Compared to [25,24], estimation of the visibility masks has been modified – an object is now considered visible at pixels with no depth measurements.”
Deep Template-based Object Instance Detection (2021, methodology): “We rely on 125 different textured 3D models gathered in majority from the various datasets of the 6D pose estimation benchmark [17] (excluding Linemod [12] since it is used for evaluation).”
Implicit 3D Orientation Learning for 6D Object Detection from RGB Images (2018, background): “Recently, Hodan et al. (2019) have shown promising results for 2D Object Detection trained on physically-based renderings.”
GDRNPP: A Geometry-Guided and Fully Learning-Based Object Pose Estimator (2021, unknown): “The BOP Challenge [94], [104] has recently become the de-facto benchmark in object pose estimation.”
CosyPose: Consistent multi-view multi-object 6D pose estimation (2020, methodology): “On T-LESS, the distance to the object is ﬁxed in the real training images and we use instead the range of distances of the testing set provided (which is explicitly allowed by the guidelines of the BOP challenge [44] 8 ).”
iNeRF: Inverting Neural Radiance Fields for Pose Estimation (2020, background): “It is a metric widely used in the pose estimation community [6].”
DexYCB: A Benchmark for Capturing Hand Grasping of Objects (2021, methodology): “Many datasets [48, 49, 16, 17] have been introduced in both domains and have facilitated progress on these two problems in parallel.”
Imbalance Problems in Object Detection: A Review (2019, background): “surveillance [1], [2], autonomous driving [3], [4], medical decision making [5], [6], and many problems in robotics [7], [8], [9], [10], [11], [12].”
FFB6D: A Full Flow Bidirectional Fusion Network for 6D Pose Estimation (2021, unknown): “The recent advent of inexpensive RGBD sensors provides extra depth information to ease the problem [5, 19, 23, 22] and also leads Pose Estimation Dense Fusion CNN Encoder CNN Decoder”