DPOD: 6D Pose Object Detector and Refiner

By Sergey Zakharov, Ivan S. Shugurov, Slobodan Ilic. IEEE International Conference on Computer Vision, 2019

498 citations recorded by the source corpus.

In this paper we present a novel deep learning method for 3D object detection and 6D pose estimation from RGB images. Our method, named DPOD (Dense Pose Object Detector), estimates dense multi-class 2D-3D correspondence maps between an input image and available 3D models. Given the correspondences, a 6DoF pose is computed via PnP and RANSAC. An additional RGB pose refinement of the initial pose estimates is performed using a custom deep learning-based refinement scheme. Our results and comparison to a vast number of related works demonstrate that a large number of correspondences is beneficial for obtaining high-quality 6D poses both before and after refinement. Unlike other methods that mainly use real data for training and do not train on synthetic renderings, we perform evaluation on both synthetic and real training data demonstrating superior results before and after refinement when compared to all recent detectors. While being precise, the presented approach is still real-time capable.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/efae64551ff0b38fb6ac938727a001a9892be67f

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 Computer Vision and Pattern Recognition

Citations in the corpus

BOP Challenge 2020 on 6D Object Localization (2020, methodology): “79 26 DPOD (synthetic) [54] 16.”
Shape Prior Deformation for Categorical 6D Object Pose and Size Estimation (2020, methodology): “[42] estimates a dense 2D-3D correspondence map between the input image and object model.”
Vision-based robotic grasping from object localization, object pose estimation to grasp estimation for parallel grippers: a review (2020, background): “DPOD [Zakharov et al., 2019] estimates dense multi-class 2D-3D correspondence maps between an input image and available 3D models.”
CosyPose: Consistent multi-view multi-object 6D pose estimation (2020, methodology): “that operate on RGB (no depth) images [7,8,10–12] are based on trainable convolutional neural networks and are able to deal with symmetric or textureless objects, which were challenging for earlier methods relying on local [3,13–16] or global [17] gradient-based image features.”
ARCTIC: A Dataset for Dexterous Bimanual Hand-Object Manipulation (2022, methodology): “• Success Rate: Following [62, 79], to measure object reconstruction quality, we use a success rate metric that is independent of the object size.”
GDRNPP: A Geometry-Guided and Fully Learning-Based Object Pose Estimator (2021, unknown): “However, the recent trend goes towards predicting dense rather than sparse correspondences, including DPOD [42], DPODv2 [26], CDPN [24], SurfEmb [43], and SDFlabel [44].”
HybridPose: 6D Object Pose Estimation Under Hybrid Representations (2020, unknown): “[44, 19] use point matching loss and achieve high accuracy.”
iNeRF: Inverting Neural Radiance Fields for Pose Estimation (2020, methodology): “Train data With Real Pose Labels Without Real Pose Labels Methods DPOD[51] DeepIM[14] AAE[39] MHP [20] DPOD[51] Self6D[48] Ours-init Ours-iNeRF Ape 53.”
DeceptionNet: Network-Driven Domain Randomization (2019, methodology): “Recent methods [21, 29, 16, 53] prerender a number of synthetic images featuring different light conditions.”