A Database for Handwritten Text Recognition Research

By J. Hull. 1994

Subfields: Computer Vision and Pattern Recognition, Evaluation & Benchmarks

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 Evaluation & Benchmarks

Citations in the corpus

Dataset Distillation (2018, methodology): “Table 2: Performance of our method and baselines in adapting models among MNIST ( M ), USPS ( U ), and SVHN ( S ).”
The Bayesian Case Model: A Generative Approach for Case-Based Reasoning and Prototype Classification (2014, methodology): “Also, LDA accuracy for the full Handwritten Digit dataset (accuracy: 0.76 ± 0.017) is comparable to that produced by BCM using the subsampled dataset (70 samples per digit, accuracy: 0.77 ± 0.03).”
Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualization (2020, methodology): “Similarly to other studies of DR methods (Tang et al., 2016; McInnes et al., 2018; Amid and Warmuth, 2019; Belkina et al., 2019), we use the following datasets to evaluate and compare DR methods: COIL-20 (Nene et al., 1996b), COIL-100 (Nene et al., 1996a), S-curve with hole dataset (synthesized by t”
Few-Shot Adversarial Domain Adaptation (2017, methodology): “We present results using the Office dataset [47], the MNIST dataset [32], the USPS dataset [24], and the SVHN dataset [40].”
Learning Task Grouping and Overlap in Multi-task Learning (2012, methodology): “Following (Argyriou et al., 2008a) • USPS Digits data: This is a handwritten digits dataset (Hull, 1994) with 10 classes.”
Unified Deep Supervised Domain Adaptation and Generalization (2017, methodology): “We present results using the Office dataset [53], the MNIST dataset [37], and the USPS dataset [31].”
Distributed optimization of deeply nested systems (2012, methodology): “The USPS dataset (Hull, 1994), a commonly used machine learning benchmark, contains 16 16 grayscale images of handwritten digits, i.e., 256D vectors with values in [0 , 1].”
Generate to Adapt: Aligning Domains Using Generative Adversarial Networks (2017, methodology): “MNIST and USPS are large datasets of handwritten digits captured under constrained conditions.”
Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation (2016, methodology): “We evaluate our algorithm’s recognition accuracy over three cross-domain pairs: 1) MNIST vs USPS, 2) SVHN vs MNIST, and 3) CIFAR vs STL. Data setup: MNIST ( mn ) vs USPS ( us ) contains 2D grayscale handwritten digit images of 10 classes.”
Learning a kernel matrix for nonlinear dimensionality reduction (2004, methodology): “The data set (Hull, 1994) consisted of N =953 grayscale images at 16×16 resolution of handwritten twos and threes (in roughly equal proportion).”
Offline Arabic handwriting recognition: a survey (2006, methodology): “For example, the CEDAR database by our group was released in 1994 and spurred intense research in the Latin OCR field [50].”
Interactive and interpretable machine learning models for human machine collaboration (2015, methodology): “I validate this through use of two standard datasets: Handwritten Digit [73] and 20 Newsgroups [88].”
See all by looking at a few: Sparse modeling for finding representative objects (2012, methodology): “The experiments are run on the USPS digits database [23] and the Extended YaleB face database [28].”
CASIA Online and Offline Chinese Handwriting Databases (2011, methodology): “Among the offline sample datasets are the CENPARMI digits, CEDAR English words and characters [1], NIST handprinted forms and characters database [2], IAM English sentence database [3], Japanese Kanji character databases ETL8B and ETL9B, Indian database of ISI [4], Arabic databases [5], Farsi databa”
A Review of Single-Source Deep Unsupervised Visual Domain Adaptation (2020, background): “USPS [157] is also a dataset of handwritten digits with 7,291 examples for training and 2,007 examples for testing.”
Combining active learning and semi-supervised learning using Gaussian fields and harmonic functions (2003, methodology): “We then evaluate active learning on a handwritten digits dataset, originally from the Cedar Buffalo binary digits database (Hull, 1994).”
A survey on off-line Cursive Word Recognition (2002, background): “g the CEDAR database [37] is composed of postal material and allows the simulation of a postal plant.”