Multidimensional Morphable Models: A Framework for Representing and Matching Object Classes

By Michael J. Jones, Tomaso A. Poggio. 2004

129 citations recorded by the source corpus.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

Learning-based Video Motion Magnification (2018, methodology): “To cope with this, we introduce another output from the encoder that represents intensity information (“texture representation” [9]) similar to the amplitude of the steerable pyramid decomposition.”
Discovering states and transformations in image collections (2015, methodology): “Several methods attempt to represent the continuous variation in an image class using subspaces [5], [22] or manifolds [13].”
Active Appearance Models Revisited (2004, methodology): “Another thing that makes empiri cal evaluation hard is the wide variety of AAM fitting algorithms [6,7,11,18,21] and the lack of a sta ndard test set.”
Face identification across different poses and illuminations with a 3D morphable model (2002, methodology): “Ignoring those points with highest contribution to EI in the fitting procedure [9] makes our algorithm robust with respect to glasses and beards.”
An expectation maximization approach to the synergy between image segmentation and object categorization (2005, methodology): “For cars we used 40 manually segmented car images from the database of [17] to learn the codebook-based detection model; since no hand-labelled keypoints were available, we used the bootstrapping algorithm in [16] with 35 of these images.”
Face Recognition Based on Fitting a 3D Morphable Model (2003, background): “Two-dimensional face models represent gray values and their image locations independently [3], [4], [18], [23], [13], [22].”
Recognizing Rotated Faces From Frontal and Side Views: An Approach Toward Effective Use of Mugshot Databases (2008, background): “A number of automatic algorithms for features location or detection have been proposed recently, such as the face vectorizing algorithm [6]; active shape model (ASM) [13], [36], [52]; Bayesian shape model (BSM) [54]; active contours [25], [29]; morphable models [33], [51]; and 2-D shape clustering [”
FAME-a flexible appearance modeling environment (2003, methodology): “[22] have used a variation of the traditional AAM, with a shape model similar to that of the Morphable Models of Jones and Poggio [23], for 3-D segmentation of the hippocampus in brain MRI.”
Birdlets: Subordinate categorization using volumetric primitives and pose-normalized appearance (2011, methodology): “Our proposed approach contributes three main innova- tions: (i) a framework, based on Poselets, for detecting volumetric part models, used both to find the basic-level object and to convey information about part shape and con- figuration; (ii) a pose-normalized appearance model (similar to represent”
Data Augmentation Using Learned Transformations for One-Shot Medical Image Segmentation (2019, methodology): “Many works build upon the framework of Morphable Models [38] or Active Appearance Models (AAMs) [15, 16], in which statistical models of shape and texture are constructed.”
Facial Landmark Detection: A Literature Survey (2018, background): “However, since the Jacobin and Hessian matrix for both p and λ need to be calculated for each iteration (Jones and Poggio 1998), the fitting procedure is usually very slow.”
Robust Object Detection with Interleaved Categorization and Segmentation (2008, background): “Prominent examples of this approach include Deformable Templates (Yuille et al. 1989; Sclaroff 1997), Morphable Models (Jones and Poggio 1998), or Shape Context Matching (Belongie et al. 2002).”