Multimodal templates for real-time detection of texture-less objects in heavily cluttered scenes

By Stefan Hinterstoißer, Stefan Holzer, Cedric Cagniart, Slobodan Ilic, Kurt Konolige, Nassir Navab, Vincent Lepetit. 2011

681 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

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

Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics (2017, background): “register the pose of 3D object models [8, 23, 53, 65] in order to index into a set of manipulation plans.”
Sim-To-Real via Sim-To-Sim: Data-Efficient Robotic Grasping via Randomized-To-Canonical Adaptation Networks (2018, neutral): “with stored objects based on visual and geometric similarity [6, 11, 19, 20, 29].”
Deep Object-Centric Representations for Generalizable Robot Learning (2017, methodology): “Instance level approaches have used methods such as 3D representations [15] [16], keypoints [17], or deep”
Learning to Rank 3D Features (2014, methodology): “Other than the local feature point based methods described in Section 1, there exists a large body of work that models 3D shapes globally such as using 2D/3D contours [3,4], shape templates [10,19], and feature histograms [33,12].”
Recovering 6D Object Pose and Predicting Next-Best-View in the Crowd (2015, methodology): “Object detection and 6 DoF pose estimation is also frequently addressed in the literature.”
Real-Time Seamless Single Shot 6D Object Pose Prediction (2017, methodology): “For example, Hinterstoisser proposed template matching algorithms suitable for both color and depth images [7, 8].”
Unsupervised Feature Learning for RGB-D Based Object Recognition (2012, background): “However, most state-of-the-art object recognition systems are still based on hand-designed representations (features), such as SIFT [25], Spin Images [17], SURF [2], Fast Point Feature Histogram [27], LINE-MOD [14], or feature combinations [19] [6].”
PointNetGPD: Detecting Grasp Configurations from Point Sets (2018, background): “Such matching is mainly based on visual and geometrical similarity [17]–[19].”
PVN3D: A Deep Point-Wise 3D Keypoints Voting Network for 6DoF Pose Estimation (2019, methodology): “LineMOD Dataset [18] consists of 13 low-textured objects in 13 videos, annotated 6D pose and instance mask.”
A codebook-free and annotation-free approach for fine-grained image categorization (2012, methodology): “Our algorithm captures the subtle differences of object classes by directly matching image regions through a highly efficient template matching algorithm inspired by [15] [14].”
DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion (2019, methodology): “In the experimental section, we would like to answer the following questions: (1) How does the dense fusion network compare to naive global fusion-by-concatenation? (2) Is the dense fusion and prediction scheme robust to heavy occlusion and segmentation errors? (3) Does the iterative refinement modu”
Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation (2019, methodology): “The OccludedLINEMOD dataset has 9 object instances and provides a CAD model for each instance.”
GDRNPP: A Geometry-Guided and Fully Learning-Based Object Pose Estimator (2021, neutral): “In the pre-deep learning era, methods can be roughly categorized into feature-based [8], [9], [10] and template-based [11], [12], [13] approaches.”
3D Object Detection with Multiple Kinects (2012, methodology): “Combination of depth and intensity has also shown to be beneﬁcial for template based method allowing real time object detection [8].”
PR2 looking at things — Ensemble learning for unstructured information processing with Markov logic networks (2014, methodology): “Subsequently, object hypotheses are annotated with a simple shape, the color, compared to known object models using the Linemod algorithm [2].”
Learning Visual Affordances for Robotic Manipulation (2019, background): “Existing frameworks such as LINEMOD [41] or MOPED [17] work well under certain assumptions such as objects sitting on a table top with good illumination, but underperform when confronted with the limited visibility, shadows, and clutter imposed by the APC scenario [18].”
CosyPose: Consistent multi-view multi-object 6D pose estimation (2020, methodology): “The object pose estimation problem [15,16] has been approached either by estimating the pose from 2D-3D correspondences using local invariant features [3,13], or directly by estimating the object pose using template-matching [14].”
SSD-6D: Making RGB-Based 3D Detection and 6D Pose Estimation Great Again (2017, neutral): “RGB Ours Brachmann 2016 [5] LineMOD [13]”
Form2Fit: Learning Shape Priors for Generalizable Assembly from Disassembly (2019, methodology): “Classic methods for assembly are often characterized by a perception module that first estimates the 6D object poses of observed parts [1], [2], [3], [4], [5], followed by a planner that then optimizes for picking and placing actions based on the inferred object poses and task-informed goal states.”