Cascaded pose regression

By Piotr Doll'ar, Peter Welinder, Pietro Perona. 2010

590 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Aerospace Engineering

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, background): “However, as is commonly done for pose regression [10] or face alignment [4], [58], [59], the regression task can be decomposed into a sequence of simpler steps.”
Cascade R-CNN: Delving Into High Quality Object Detection (2017, methodology): “The difficult regression task can be decomposed into a sequence of simpler steps, inspired by the works of cascade pose regression [6] and face alignment [2, 35].”
End-to-End Recovery of Human Shape and Pose (2017, methodology): “In this work, we take inspiration from previous works [7, 9, 31] and regress Θ in an iterative error feedback (IEF) loop, where progressive changes are made recurrently to the current estimate.”
Human Pose Estimation with Iterative Error Feedback (2015, methodology): “The system in our human pose estimation example can be seen as closest to approaches employing “pose-indexed features” [12, 8, 17], but leveraging hierarchical feature learning.”
Incremental Face Alignment in the Wild (2014, background): “Similar ideas have been explored for human pose estimation [9].”
FSA-Net: Learning Fine-Grained Structure Aggregation for Head Pose Estimation From a Single Image (2019, methodology): “Regression-based methods [5, 43, 12, 23, 42] sketch initial faces, and incrementally align the drawn faces to real ones by regression.”
Coarse-to-Fine Auto-Encoder Networks (CFAN) for Real-Time Face Alignment (2014, methodology): “Moreover, in each stage of the cascades architecture of SDM, linear regression is exploited to model the mapping from shape-index feature [12] to face shape, which may be insufficient for the complex non-linear process from shape-index feature to face shape.”
Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network (2018, methodology): “lignment approaches which aim at locating a set of ducial 2D facial landmarks, such as classic Active Appearance Model(AMM)[40,41,42] and Constrained Local Models(CLM)[43,44]. Then cascaded regression[45,46] and CNNbased methods[3,4,19] are largely used to achieve state-of-the-art performance in 2D ”
Mnemonic Descent Method: A Recurrent Process Applied for End-to-End Face Alignment (2016, methodology): “Cascaded regression methodologies have been also proposed in the recent works of [20, 11, 4, 49, 37, 26, 50].”
One millisecond face alignment with an ensemble of regression trees (2014, background): “Recent work [8, 11, 2] uses the fact that a certain class of regressors are guaranteed to produce predictions that lie in a linear subspace defined by the training shapes and there is no need for additional constraints.”
Convolutional Channel Features (2015, unknown): “Besides visual recognition task (e.g. pedestrian detection), channel features are also successfully applied to other vision tasks like pose estimation [13], edge detection [15] and object proposal generation [56].”
Pose Recognition with Cascade Transformers (2021, methodology): “Furthermore, regression-based approaches typically have a recursive procedure [9] that skips a large number of candidate locations, creating a performance gap with the heatmap-based methods.”
Cascaded hand pose regression (2015, methodology): “While we use the similar pixel difference features as in cascaded pose regression [9], face [5], human body [18, 29] and hand [6, 39, 10, 20, 22], we show that the pixel parameterization is the key to achieve certain geometrical invariance.”
Face Alignment by Explicit Shape Regression (2012, methodology): “Such features are extremely cheap to compute and powerful enough given sufficient training data [15, 18, 7].”
Robust 3D Hand Pose Estimation in Single Depth Images: From Single-View CNN to Multi-View CNNs (2016, methodology): “for complex hand postures, pre-learned hand pose priors are sometimes utilized to better fuse independently predicted hand joint distributions [8, 12]. In [21], the cascaded pose regression algorithm [3] is adapted to the problem of hand pose estimation. Particularly, the authors propose to ﬁrst pre”
Benchmarking and Error Diagnosis in Multi-instance Pose Estimation (2017, background): “Pose Estimation: In their early work on pose regression, Doll´ar et al. [13] observed that unlike human annotators, algorithms have a distribution of the normalized distances between a part detection and its ground-truth that is typically bimodal, highlighting the presence of multiple error modes.”
Face alignment by coarse-to-fine shape searching (2015, methodology): “Face alignment by cascaded regression: There are a few successful methods that adopt the concept of cascaded pose regression [17].”