Robust Face Landmark Estimation under Occlusion

By Xavier Paolo Burgos-Artizzu, Pietro Perona, Piotr Doll'ar. IEEE International Conference on Computer Vision, 2013

814 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/2724ba85ec4a66de18da33925e537f3902f21249

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 Signal Processing

Citations in the corpus

Unsupervised Learning of Object Landmarks through Conditional Image Generation (2018, methodology): “Method K MAFL AFLW Supervised RCPR [2] – 11.”
Deep High-Resolution Representation Learning for Visual Recognition (2019, methodology): “The COFW dataset [10] consists of 1 , 345 training and 507 testing faces with occlusions, where each image has 29 facial landmarks.”
Unsupervised Discovery of Object Landmarks as Structural Representations (2018, background): “2, 4 [5] Xavier P Burgos-Artizzu, Pietro Perona, and Piotr Dollár.”
HyperFace: A Deep Multi-Task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition (2019, methodology): “AAM) [70], Constrained Local Model (CLM) [45], Oxford facial landmark detector [11], Zhu [70], FaceDPL [71], JointCascade [6], CDM [62], RCPR [3], ESR [4], SDM [56] and 3DDFA [69].”
Mnemonic Descent Method: A Recurrent Process Applied for End-to-End Face Alignment (2016, methodology): “Cascaded regression strategies constitute a large portion of the most popular facial alignment algorithms, as they are highly efficient and generalise well [56, 37, 11, 10, 3, 26, 31, 64, 57, 50].”
Adaptive Wing Loss for Robust Face Alignment via Heatmap Regression (2019, methodology): “Our approach outperforms the state-of-the-art algorithms by a significant margin on mainstream face alignment datasets including 300W [53], COFW [8] and WFLW [62].”
Face Alignment at 3000 FPS via Regressing Local Binary Features (2014, methodology): “Our work belongs to the shape regression approach [5, 11, 12, 29, 4, 32, 3] category.”
On Face Segmentation, Face Swapping, and Face Perception (2017, methodology): “They report state of the art performance on the Caltech Occluded Faces in the Wild (COFW) dataset [6].”
A Comprehensive Analysis of Deep Regression (2018, neutral): “In computer vision, regression techniques span a large ensemble of applicative scenarios such as: head-pose estimation [1], [2], facial landmark detection [3], [4], human pose estimation [5], [6], age estimation [7], [8], or image registration [9], [10].”
Learning Deep Representation for Face Alignment with Auxiliary Attributes (2014, methodology): “Index Terms—Face Alignment, face landmark detection, deep learning, convolutional network Ç”
OpenFace 2.0: Facial Behavior Analysis Toolkit (2018, neutral): “This leads to improved accuracy for facial landmark detection, head COFW[13] RCPR[13] FaceTracker CLM[57] dlib [37] [35] Chehra [5] Menpo [2] AAM, CLM, SDM [85] [85] OpenPose [15] Part afﬁnity Fields [15] 3 CFSS [83] CFSS [83] iCCR [56] iCCR [ Our work is intended to bridge that gap between…”
An All-In-One Convolutional Neural Network for Face Analysis (2016, methodology): “7(a) shows comparisons with recently published methods such as CCL [57], HyperFace [36], LBF [38], SDM [47], ERT [22] and RCPR [1].”
Wing Loss for Robust Facial Landmark Localisation with Convolutional Neural Networks (2017, methodology): “We compare our method with a set of state-of-the-art approaches, including SDM [63], ERT [30], RCPR [5], CFSS [74], LBF [46], GRF [23], CCL [75], DAC-CSR [21] and TR-DRN [39].”
Facial Landmark Detection: A Literature Survey (2018, background): “12 Facial images (Burgos-Artizzu et al. 2013)with different occlusions”
UMDFaces: An annotated face dataset for training deep networks (2016, background): “The NME(%) of face alignment results on AFLW test set for various poses (frontal ([0-30]) to profile ([60-90])).”
High-Resolution Representations for Labeling Pixels and Regions (2019, methodology): “We perform the evaluation over four standard datasets: WFLW [101], AFLW [49], COFW [8], and 300 W [85].”