Training set size requirements for the classification of a specific class

By Giles M. Foody, Ajay Mathur, C. Sanchez-Hernandez, D. Boyd. 2006

320 citations recorded by the source corpus.

Subfields: Analytical Chemistry, Ecology, General Agricultural and Biological Sciences

View Paper: https://www.semanticscholar.org/paper/900a9cd2312da71d62dc4f6fe405cb8ce3a6b3e4

How researchers see this paper

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

In Analytical Chemistry

In Ecology

In General Agricultural and Biological Sciences

Giles M. Foody

Citations in the corpus

Measuring and modelling biodiversity from space (2008, background): “Franklin and Wulder (2002) undertook an excellent review of large-scale land-cover classifi cations, such as CORINE and GAP, that generally seek to attain 85% accuracy across all mapping classes using a variety of passive sensors (TM, SPOT, AVHRR, MODIS) and to a lesser extent active sensors…”
Accurate Annotation of Remote Sensing Images via Active Spectral Clustering with Little Expert Knowledge (2015, background): “Highly redundant information and noise in labeled training data may lead to poor performance [23].”
Comparing Machine Learning Classifiers for Object-Based Land Cover Classification Using Very High Resolution Imagery (2014, background): “There are some basic principles for the selection of training samples for pixel-based classification [18,19].”
Multiclass and Binary SVM Classification: Implications for Training and Classification Users (2008, background): “Moreover, it may sometimes be possible to identify the most useful training sites for the provision of support vectors before the classification [3], [4].”
Operational High Resolution Land Cover Map Production at the Country Scale Using Satellite Image Time Series (2017, background): “Although some studies about the ad-hoc selection of samples in order to reduce training set sizes exist [9,10], the impact of the scarcity and spatial distribution of reference data for the mapping of large areas using supervised classification has not been assessed in the literature.”
Support vector machine to map oil palm in a heterogeneous environment (2014, methodology): “The multiclass SVM approach with an RBF kernel was used to classify the data (Huang, Davis, and Townshend 2002; Foody et al. 2006; Foody and Mathur 2004a).”
Comparison of Classification Algorithms and Training Sample Sizes in Urban Land Classification with Landsat Thematic Mapper Imagery (2014, background): “Although more training samples are usually beneficial, as they tend to be more representative to the class population, a small number of training samples is obviously attractive for logistic reasons [14].”
A Survey of Active Learning Algorithms for Supervised Remote Sensing Image Classification (2011, background): “Moreover, the inclusion of noisy pixels may result in a wrong representation of the class statistics, which may lead to poor classification performances and/or overfitting [13].”
Support vector machines in remote sensing: A review (2011, unknown): “Foody and Mathur (2006) proposed a focus on mixed pixel training samples over more tedious, conventional pure pixel acquisition, assuming an SVM classifier.”