Metrics and Models for Handwritten Character Recognition

By Trevor J. Hastie, Patrice Y. Simard. 1998

How researchers see this paper

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

Citations in the corpus

Prototype selection for interpretable classification (2011, methodology): “Use of tangent distance with 1-NN attained the lowest test errors of any method [Hastie and Simard (1998)].”
Transformation Invariance in Pattern Recognition-Tangent Distance and Tangent Propagation (1996, background): “It is also straightforward to derive an batch [13] or on-line [23] algorithm to train the tangent vectors”
Nonlocal Estimation of Manifold Structure (2006, background): “Hastie, Simard, and Sackinger (1995) and Hastie and Simard (1998) present a tangent subspace learning algorithm to learn character prototypes along with a tangent plane around each prototype, which reduces the time and memory requirements of the nearest-neighbor tangent distance classifier.”
Sparse Subspace Clustering: Algorithm, Theory, and Applications (2012, methodology): “We demonstrate the effectiveness of the proposed algorithm through experiments on synthetic data as well as the two real-world problems of motion segmentation and face clustering.”
Transformation invariance in pattern recognition: Tangent distance and propagation (2000, background): “The derivation has been done for batch training (Hastie and Simard, 1998) and for on-line training (Schwenk, 1998) of the tangent vectors.”
Multi-Scale Fusion Subspace Clustering Using Similarity Constraint (2020, background): “For example, the face images of a subject taken under ﬁxed pose, varying lighting conditions and Lambertian reﬂectance, occupy a low-dimension subspace whose dimension close to nine [2, 13], and the handwritten digit images of a single digit also form a low-dimension subspace [10].”
Sparse Subspace Clustering: Algorithm, Theory, and Applications (2013, background): “For example, feature trajectories of a rigidly moving object in a video [2], face images of a subject under varying illumination [3], and multiple instances of a hand-written digit with different rotations, translations, and thicknesses [4] lie in a low-dimensional subspace of the ambient space.”
A Geometric Analysis of Subspace Clustering with Outliers (2011, background): “[17], these 7-dimensional manifolds are in turn well approximated by 12-dimensional subspaces.”
Subspace Clustering by Block Diagonal Representation (2018, methodology): “ing images of handwritten digits which also have the subspace structure of dimension 12 [13].”