2D Human Pose Estimation: New Benchmark and State of the Art Analysis

By Mykhaylo Andriluka, Leonid Pishchulin, Peter Gehler, Bernt Schiele. 2014

2,802 citations recorded by the source corpus.

Human pose estimation has made significant progress during the last years. However current datasets are limited in their coverage of the overall pose estimation challenges. Still these serve as the common sources to evaluate, train and compare different models on. In this paper we introduce a novel benchmark "MPII Human Pose" that makes a significant advance in terms of diversity and difficulty, a contribution that we feel is required for future developments in human body models. This comprehensive dataset was collected using an established taxonomy of over 800 human activities [1]. The collected images cover a wider variety of human activities than previous datasets including various recreational, occupational and householding activities, and capture people from a wider range of viewpoints. We provide a rich set of labels including positions of body joints, full 3D torso and head orientation, occlusion labels for joints and body parts, and activity labels. For each image we provide adjacent video frames to facilitate the use of motion information. Given these rich annotations we perform a detailed analysis of leading human pose estimation approaches and gaining insights for the success and failures of these methods.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

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

In Biomedical Engineering

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

Mask R-CNN (2017, methodology): “†: G-RMI was trained on COCO plus MPII [1] (25k images), using two models (Inception-ResNet-v2 + ResNet101).”
Towards Accurate Multi-person Pose Estimation in the Wild (2017, methodology): “Single-Person Pose With the development of Deep Convolutional Neural Networks (CNN) for vision tasks, state-of-art performance on pose estimation is achieved using CNNs [43, 26, 42, 10, 31, 2, 7, 6, 20, 24, 8].”
Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields (2016, methodology): “We evaluate our method on two benchmarks for multi-person pose estimation: (1) the MPII human multi-person dataset [2] and (2) the MSCOCO 2016 keypoints challenge dataset [21].”
Convolutional Pose Machines (2016, background): “The evaluation is based on PCKh metric [1] where the error tolerance is normalized with respect to head size of the target.”
DensePose: Dense Human Pose Estimation in the Wild (2018, background): “Our approach can be understood as the next step in the line of works on human pose estimation [27, 1, 20, 7, 42, 19, 29].”
Extreme Clicking for Efficient Object Annotation (2017, background): “A qualification test is a good mechanism for enhancing the quality of crowd-sourcing data and for filtering out bad annotators and spammers [1, 19, 33, 37].”
The Kinetics Human Action Video Dataset (2017, methodology): “These sources include: (i) Action datasets – existing datasets like Ac-tivityNet [3], HMDB [15], UCF101 [20], MPII Human Pose [2], ACT [25] have useful classes and a suitable sub set of these were used; (ii) Motion capture – there are a number of motion capture datasets which we looked through and…”
Parallel Multiscale Autoregressive Density Estimation (2017, background): “We evaluate our model on ImageNet, Caltech-UCSD Birds (CUB), the MPII Human Pose dataset (MPII), the Microsoft Common Objects in Context dataset (MS-COCO), and the Google Robot Pushing dataset.”
Learning What and Where to Draw (2016, background): “This suggests that the model learns to use the the noise distribution to capture some aspects of the background and also non-controllable aspects of “where” such as direction.”
Attention to Scale: Scale-Aware Semantic Image Segmentation (2015, methodology): “Supplementary materials: In the supplementary materials, we apply our trained model to some videos from MPII Human Pose dataset [1].”
Multimodal Explanations: Justifying Decisions and Pointing to the Evidence (2018, methodology): “Additionally, we collect annotations for activity recognition using the MPII Human Pose (MHP) dataset [2].”
Feature Boosting Network For 3D Pose Estimation (2019, methodology): “6M and MPII are used for training, while the testing is performed on MPI-INF-3DHP.”
2D/3D Pose Estimation and Action Recognition Using Multitask Deep Learning (2018, methodology): “For training the pose estimation part, we use mixed data from MPII (75”
Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection (2017, methodology): “We trained the Fast-Net architecture on the J-HMDB [13] and the MPII [1] action recognition datasets.”
End-to-End Recovery of Human Shape and Pose (2017, methodology): “The λs and other hyper-parameters are set through validation data on MPIINF-3DHP dataset.”
Searching for Efficient Multi-Scale Architectures for Dense Image Prediction (2018, neutral): “74% improvement over the best state-of-the-art model [25], consistently outperforming [1], as required in [89, 25].”
Human Pose Estimation with Iterative Error Feedback (2015, methodology): “We tested our method on the two most challenging benchmarks for 2D human pose estimation: the MPII Human Pose dataset [2], which features significant scale variation, occlusion, and multiple people interacting, and Leeds Sports Pose dataset (LSP) [19] which features complex poses of people in sports”