View-based active appearance models

By Tim Cootes, Gavin V. Wheeler, Kevin N. Walker, Christopher J. Taylor. Proceedings Fourth IEEE International Conference on Automatic Face and Gesture Recognition (Cat. No. PR00580), 2000

623 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/ee22010e4e8139df9c2e62944759837b48f58cae

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

Citations in the corpus

Face Alignment in Full Pose Range: A 3D Total Solution (2015, methodology): “The most common method is the multi-view framework [14], which uses different landmark configurations for different views.”
Real Time Head Pose Estimation from Consumer Depth Cameras (2011, background): “, active appearance models [8] and their variations [7, 19, 2], are very popular in the face analysis field, but are rarely employed for head pose estimation.”
Computer Vision - ECCV 2004 (2004, methodology): “parameters (M-step) and suffer from the disadvantage that the convergence to the optimal solution strongly depends on correct initialization [5,6].”
Evaluation of prostate segmentation algorithms for MRI: the PROMISE12 challenge (2013, background): “Examples included coupled active appearance models (Cootes et al., 2000) or hierarchical segmentation strategies (Wolz et al., 2012).”
Face recognition across pose: A review (2009, methodology): “The texture mapping is similar to AAM, except that the model parameters of texture model are independent to shape model.”
Recognizing Rotated Faces From Frontal and Side Views: An Approach Toward Effective Use of Mugshot Databases (2008, methodology): “Given the current state of technologies, however, computerized face recognition requires cooperative subjects who stay still in Manuscript received October 09, 2007; revised April 21, 2008.”
Facial Landmark Detection: A Literature Survey (2018, methodology): “2003; Cootes et al. 2000; Zhu and Ramanan 2012; Dantone et al. 2012). They either select the best model or merge the results from all models. There are two ways to select the model. The first way is to estimate the head poses using existing head pose estimation methods. For example, in the early wor”
Face recognition: A literature survey (2003, methodology): “For example, a view-based statistical method claims to be able to handle even pro.le views in which many local features are invisible [Cootes et al. 2000].”
Automatic Sign Language Analysis: A Survey and the Future beyond Lexical Meaning (2005, methodology): “They used a network of 2D Active Appearance Models [28] as an intermediate representation between image features, and a simplified 3D hand model with 9 degrees-of-freedom.”
Describable Visual Attributes for Face Verification and Image Search (2011, neutral): “The CMU Pose, Illumination, and Expression (PIE) data set and follow-on results showed that sometimes alignment, especially in 3D, can overcome the other difficulties [4], [7], [10], [24], [52].”
From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose (2001, methodology): “In another approach, an Active Appearance Model of a generic face is deformed to fit to the input image and the control parameters are used as a feature vector for classification [10], [39], [14], [11].”
Face identification across different poses and illuminations with a 3D morphable model (2002, neutral): “Head poses that range from frontal to profile views need to be covered in view-based methods by a set of separate models for different views [3].”
Random Forests for Real Time 3D Face Analysis (2012, methodology): “A very popular family of methods use statistical models of the face shape and appearance, like Active Appearance Models (AAMs) (Cootes et al. 2001), multi-view AAMs (Ramnath et al. 2008), and 3D Morphable Models (Blanz and Vetter 1999; Storer et al. 2009).”
A Survey on Human Motion Analysis from Depth Data (2013, background): “Holistic methods, e.g., Active Appearance Models [49, 71, 72], use the entire facial texture to fit a generative model to a test image.”
Inferring 3D structure with a statistical image-based shape model (2003, background): “Shape models in several views can be separately estimated to match object appearance [5]; this approach was able to learn a mapping between the low-dimensional shape parameters in each view.”
3D Constrained Local Model for rigid and non-rigid facial tracking (2012, methodology): “There have been several extensions to Active Appearance Models that explicitly model the 3D shape in the formulation of the PDM [29], or train several types of models for different view points [8].”
Locally Linear Regression for Pose-Invariant Face Recognition (2007, background): “By extending it to view-based AAM, novel views of the input face can be synthesized and used for pose-invariant face recognition [20].”
Real-time facial feature detection using conditional regression forests (2012, background): “, Active Appearance Models [2, 5, 7], use the texture over the whole face region to fit a linear generative model to a test image.”
Face detection, pose estimation, and landmark localization in the wild (2012, methodology): “Most methods use explicit 3D models [6, 17] or 2D view-based models [30, 10, 21].”