Real-Time Continuous Pose Recovery of Human Hands Using Convolutional Networks

By Jonathan Tompson, Murphy Stein, Yann LeCun, Ken Perlin. 2014

824 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering, Human-Computer Interaction

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

In Human-Computer Interaction

Citations in the corpus

Learning from Simulated and Unsupervised Images through Adversarial Training (2016, methodology): “Example refined test images for the NYU hand pose dataset [38].”
Joint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation (2014, background): “The best performing algorithms today for many vision tasks in general [38, 28], and human pose estimation in particular ([36, 16, 35]) are based on deep convolutional networks.”
DISCO Nets : DISsimilarity COefficients Networks (2016, background): “[27] to estimate the efficiency of DISCO Nets for this task.”
Feature Boosting Network For 3D Pose Estimation (2019, methodology): “Different from the NYU Hand Pose dataset [33], which is mainly designed for hand pose estimation from depth images and the registered color images contain lots of artifacts, the large 3DHandPose dataset is highly suitable for 3D hand pose estimation from a single RGB image, as analyzed in [44].”
Structural-RNN: Deep Learning on Spatio-Temporal Graphs (2015, positive): “[58] jointly train CNN and MRF for human pose estimation.”
Dense Human Body Correspondences Using Convolutional Networks (2015, background): “Popular skeleton extraction methods from single-view 3D captures such as [38, 46, 43] often have difficulties resolving this ambiguity.”
SurfNet: Generating 3D Shape Surfaces Using Deep Residual Networks (2017, methodology): “Unlike standard methods which estimate the joint angle [7, 29] or joint position parameters [27, 32, 33], we reconstruct the full 3D surface.”
Self-Supervised Visual Descriptor Learning for Dense Correspondence (2017, methodology): “a set of canonical keypoints as in [25].”
Two-hand Global 3D Pose Estimation using Monocular RGB (2020, background): “For depth-based hand pose estimation, [28, 26, 25, 34, 7] presented large-scale datasets consisting of real depth images with estimated ground-truth joint locations.”
Hand Keypoint Detection in Single Images Using Multiview Bootstrapping (2017, background): “These are broadly classifiable into generative methods [14], discriminative methods [24, 25, 9, 30, 22, 27], or hybrid methods [20, 16, 18, 26, 31].”
3D Hand Shape and Pose From Images in the Wild (2019, background): “There is a rich literature on 3D hand pose and reconstruction from depth [62, 66, 10, 36, 61, 11, 43, 45, 19, 20, 24, 30, 37, 48, 52, 53, 59, 64], image and depth [26, 32, 49, 28], stereo [33, 65, 41] and multiple images [8, 31, 50].”
Spatial Attention Deep Net with Partial PSO for Hierarchical Hybrid Hand Pose Estimation (2016, methodology): “On NYU dataset, our method performs only slightly better than the best result of HSO.”
Hand Pose Estimation through Weakly-Supervised Learning of a Rich Intermediate Representation (2015, methodology): “While more recent work on body and hand pose estimation tends to perform direct regression from depth or color input to joint positions [32, 31, 3, 23], we argue that the intermediate representation is a powerful tool in the special context of training with multiple hetero geneous training sets.”
WorkingHands: A Hand-Tool Assembly Dataset for Image Segmentation and Activity Mining (2019, background): “In addition, we include depth images to provide extra information, since depth has been shown to be useful for understanding human body parts [30, 32].”
DexPilot: Vision-Based Teleoperation of Dexterous Robotic Hand-Arm System (2019, background): “Although various datasets [9, 10] exist that provide the necessary annotations for hand pose and segmentation, these datasets were not suitable primarily due to different sensor characteristics and lack of hand poses that were necessary”
TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 1 ModDrop (2018, neutral): “A group of recent works is dedicated to inferring the hand pose through pixel-wise hand segmentation and estimating the positions of hand or body joints in a bottom-up fashion [20], [21], [22], [23], [24].”
Depth-Based Hand Pose Estimation: Methods, Data, and Challenges (2015, neutral): “Others use a geometric model for inverse-kinematic (IK) refinement/validation during a “post-processing” stage [17,33,38, 42].”
Learning to Navigate the Energy Landscape (2016, methodology): “These models have led to the development of hybrid architectures, where discriminatively trained feed-forward predictors (RFs or CNNs) are used to initial-1 ize continuous local search algorithms.”