Hierarchical graphical-based human pose estimation via local multi-resolution convolutional neural network
Source abstract
Abstract
This paper addresses the problems of the graphical-based human pose estimation in still images, including the diversity of appearances and confounding background clutter. We present a new architecture for estimating human pose using a Convolutional Neural Network (CNN). Firstly, a Relative Mixture Deformable Model (RMDM) is defined by each pair of connected parts to compute the relative spatial information in the graphical model. Secondly, a Local Multi-Resolution Convolutional Neural Network (LMR-CNN) is proposed to train and learn the multi-scale representation of each body parts by combining different levels of part context. Thirdly, a LMR-CNN based hierarchical model is defined to explore the context information of limb parts. Finally, the experimental results demonstrate the effectiveness of the proposed deep learning approach for human pose estimation.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Human Pose and Action Recognition · Hand Gesture Recognition Systems · Video Surveillance and Tracking Methods
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Computer Science and Technology, Nanjing Tech University 1 , 211816 Nanjing, China
School of Automation Science and Electrical Engineering, Beihang University 2 , 100083 Beijing, China
ICD - LM2S, Université de Technologie de Troyes 3 , 10004 Troyes Cedex, France
Funding
National Natural Science Foundation of China · Aeronautical Science Foundation of China
References (54)
Show source-provided references
1. Tied factor analysis for unconstrained face pose classification10.1016/j.ijleo.2016.09.048 ↗
2. Visual pose measurement based on structured light for MAVs in non-cooperative environments10.1016/j.ijleo.2015.09.041 ↗
3. Multi-pose face ensemble classification aided by Gabor features and deep belief nets10.1016/j.ijleo.2015.10.179 ↗
4. Human pose estimation with multiple mixture parts model based on upper body categories10.1117/1.jei.24.4.043021 ↗
5. S. A. Johnson, “Articulated human pose estimation in natural images,” Ph.D. thesis, University of Leeds (2012).
6. Object detection with discriminatively trained part based models10.1109/tpami.2009.167 ↗
7. MODEC: Multimodal decomposable models for human pose estimation
8. Articulated pose estimation with flexible mixtures-of-parts
9. Pictorial structures for object recognition10.1023/b:visi.0000042934.15159.49 ↗
10. A multi-layer composite model for human pose estimation
11. Appearance sharing for collective human pose estimation
12. Global pose estimation using non-tree models
13. Fast globally optimal 2D human detection with loopy graph models
14. Strong appearance and expressive spatial models for human pose estimation
15. Human pose estimation with fields of parts
16. Object recognition from local scale-invariant features
17. Histograms of oriented gradients for human detection
18. Deep neural networks for object detection
19. Overfeat: Integrated recognition, localization and detection using convolutional networks
20. Rich feature hierarchies for accurate object detection and semantic segmentation
21. Part-based R-CNNS for fine-grained category detection
22. Spatial pyramid pooling in deep convolutional networks for visual recognition
23. Deep convolution neural networks in computer vision10.5573/ieiespc.2015.4.1.035 ↗
24. Salient object detection based on amplitude spectrum optimization
25. Gaze-assisted multi-stream deep neural network for action recognition
26. Image quality assessment using deep convolutional networks10.1063/1.5010804 ↗
27. A limb-based graphical model for human pose estimation10.1109/tsmc.2016.2639788 ↗
28. Deeppose: Human pose estimation via deep neural networks
29. Combining local appearance and holistic view: Dual-source deep neural networks for human pose estimation
30. Beyond physical connections: Tree models in human pose estimation
31. Beyond tree structure models: A new occlusion aware graphical model for human pose estimation
32. A discriminatively trained, multiscale, deformable part model
33. Monocular 3D pose estimation and tracking by detection
34. Cascaded models for articulated pose estimation
35. Detecting people using mutually consistent poselet activations
36. Learning hierarchical poselets for human parsing
37. Articulated part-based model for joint object detection and pose estimation
38. Exploring the spatial hierarchy of mixture models for human pose estimation
39. Poselet conditioned pictorial structures
40. Pose machines: Articulated pose estimation via inference machines
41. Articulated pose estimation by a graphical model with image dependent pairwise relations
42. Real-time continuous pose recovery of human hands using convolutional networks10.1145/2629500 ↗
43. Learning human pose estimation features with convolutional networks
44. Joint training of a convolutional network and a graphical model for human pose estimation
45. Selective search for object recognition
46. Fast R-CNN
47. Faster R-CNN: Towards real-time object detection with region proposal networks10.1109/tpami.2016.2577031 ↗
48. Better appearance models for pictorial structures
49. Progressive search space reduction for human pose estimation
50. Improved human parsing with a full relational model
51. Clustered pose and nonlinear appearance models for human pose estimation
52. Efficient discriminative learning of parts-based models
53. Articulated people detection and pose estimation: Reshaping the future
54. Caffe: Convolutional architecture for fast feature embedding