Learning Joint Top-Down and Bottom-up Processes for 3D Visual Inference

By Cristian Sminchisescu, Atul Kanaujia, Dimitris N. Metaxas. 2006

118 citations recorded by the source corpus.

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

VNect (2017, background): “2006] or in combination with discriminative re-initialization [Rosales and Sclaroff 2006; Sminchisescu et al. 2006], since generative reconstruction is fundamentally underconstrained.”
Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes: The Importance of Multiple Scene Constraints (2018, background): “Systems based on discriminative-generative (feedforward-feedback) components for 3d human pose estimation date, in principle, back at least to [25, 28, 31] but our approach leverages considerably different image representations, body models, cost functions and optimization techniques.”
Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments (2014, background): “Prior effors to create mixed reality setups for training and testing 3D human pose estimation methods exist, including our own prior work [41], [42] but also [43] and more recently [44].”
Sim2real transfer learning for 3D human pose estimation: motion to the rescue (2019, background): “Principal among these is work on using datasets of synthetic humans for human pose estimation [11, 16, 20, 51, 67, 75, 84, 86, 97].”
Pictorial Human Spaces: How Well Do Humans Perceive a 3D Articulated Pose? (2013, background): “An interesting avenue not pursued in almost any artificial recognition system, but not inconsistent with our findings, would be the combination of lowresolution (currently pervasive computer vision) inference with pose and image dependent search strategies that focus on high resolution features–comb”
Conditional models for contextual human motion recognition (2006, background): “of models with conditional and generative subcomponents (CRF/MEMMs and HMMs) is likely to be a promising avenue for future research [ 45 ].”
Deep Multitask Architecture for Integrated 2D and 3D Human Sensing (2017, methodology): “There is also interest in combining 2d and 3d estimation methods in order to obtain models capable of multiple task, and able to operate in realistic imaging conditions [41, 32, 47, 2].”
Learning to Reconstruct People in Clothing From a Single RGB Camera (2019, background): “A recent trend of works combines bottom-up and top-down approaches–a combination that has been exploited already in earlier works [66].”
Vision-based human motion analysis: An overview (2007, background): “[103] learn top–down and bottom–up functions in alternate steps.”
Monocular 3D Human Pose Estimation Using Transfer Learning and Improved CNN Supervision (2016, background): “Combinig discriminative and generative methods yields increased robustness outdoors [70, 67, 3, 8, 21, 55, 56, 11].”
Learning from Synthetic Humans (2017, background): “ynthetically generated videos of cars to study multi-object tracking. Several works focused on creating synthetic images of human bodies for learning 2D pose estimation [27,30,36], 3D pose estimation [7,9,14,24,35,39,44], pedestrian detection [22,27,28], and action recognition [31,32]. Pishchulin et”