ArtTrack: Articulated Multi-Person Tracking in the Wild

By Eldar Insafutdinov, Mykhaylo Andriluka, Leonid Pishchulin, Siyu Tang, Evgeny Levinkov, Bjoern Andres, Bernt Schiele. 2016

279 citations recorded by the source corpus.

In this paper we propose an approach for articulated tracking of multiple people in unconstrained videos. Our starting point is a model that resembles existing architectures for single-frame pose estimation but is substantially faster. We achieve this in two ways: (1) by simplifying and sparsifying the body-part relationship graph and leveraging recent methods for faster inference, and (2) by offloading a substantial share of computation onto a feed-forward convolutional architecture that is able to detect and associate body joints of the same person even in clutter. We use this model to generate proposals for body joint locations and formulate articulated tracking as spatio-temporal grouping of such proposals. This allows to jointly solve the association problem for all people in the scene by propagating evidence from strong detections through time and enforcing constraints that each proposal can be assigned to one person only. We report results on a public MPII Human Pose benchmark and on a new MPII Video Pose dataset of image sequences with multiple people. We demonstrate that our model achieves state-of-the-art results while using only a fraction of time and is able to leverage temporal information to improve state-of-the-art for crowded scenes.

Subfields: Computer Vision and Pattern Recognition, 3D Vision & Pose

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 3D Vision & Pose

Citations in the corpus

Deep Affinity Network for Multiple Object Tracking (2018, background): “pre-trained for classification tasks are popular [57], [58], [59], [60], [61].”
Single-Shot Multi-person 3D Pose Estimation from Monocular RGB (2017, background): “Hence, a body of work ﬁrst localizes the joints of each person with CNN-based detectors and then ﬁnd the correct association between joints and subjects in a post-processing step [43, 16, 36, 8].”
Monocular human pose estimation: A survey of deep learning-based methods (2020, methodology): “and others have pose annotations for multiple people (Insafutdinov et al., 2017; Iqbal et al., 2016; Andriluka et al., 2018).”
Neural Body Fitting: Unifying Deep Learning and Model Based Human Pose and Shape Estimation (2018, background): “The progress in 2D pose estimation using CNNs [65, 7, 15] has contributed significantly to the task of 3D pose estimation even in challenging in-the-wild scenarios.”
OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields (2018, unknown): “[47] further simplified their bodypart relationship graph for faster inference in single-frame model and formulated articulated human tracking as spatiotemporal grouping of part proposals.”
Human Pose Regression by Combining Indirect Part Detection and Contextual Information (2017, unknown): “In what follows, we ﬁrst present the global architecture of our method, and then detail its most important parts.”
Multiple object tracking: A literature review (2014, background): “Possible combinations include object segmentation [206, 207, 208, 209], re-identification [210, 194, 211], human pose estimation [18, 212, 213, 214, 215], and action recognition [19].”
Ordinal Depth Supervision for 3D Human Pose Estimation (2018, background): “Leveraging large-scale datasets with extensive 2D annotations has immensely benefited 2D pose estimation [55, 34, 31, 61], semantic part labeling [11, 57] and multi-person pose estimation [15, 30, 9].”
Deep learning tools for the measurement of animal behavior in neuroscience (2019, background): “(30) the work was extended to fast multi-human pose estimation in videos.”
Survey on Emotional Body Gesture Recognition (2018, background): “One recent example of multi-person pose estimation and tracking is Arttrack [44].”
Video Based Reconstruction of 3D People Models (2018, background): “The advance in 2D pose detection [70, 11, 32] has made 3D pose and shape estimation possible in challenging scenarios.”
Deep Learning for Person Re-Identification: A Survey and Outlook (2020, background): “The bounding boxes are usually obtained by the person detection [40], [41] or tracking algorithms [42], [43].”
Pose Flow: Efficient Online Pose Tracking (2018, methodology): “PoseTrack [13] and ArtTrack [12] in CVPR’17 primarily introduce multi-person pose tracking challenge and propose a new graph partitioning formulation, building upon 2D DeeperCut [11] by extending spatial joint graph to spatio-temporal graph.”
PoseTrack: A Benchmark for Human Pose Estimation and Tracking (2017, methodology): “In the following we first describe our baselines based on [17, 22] and then summarize the main observations made in this evaluation.”
Panoptic Studio: A Massively Multiview System for Social Interaction Capture (2016, background): “Recently, several methods have been proposed that can detect multiple people in a single image [43], [44] or a video [45].”
Deep Learning-based Human Pose Estimation: A Survey (2020, background): “, [17, 32, 57, 76, 89, 105, 171, 176, 229, 245, 253]) has two main steps including body joint detection (i.”