An Agreement Coefficient for Image Comparison

By Lei Ji, K. Gallo. 2006

130 citations recorded by the source corpus.

Combination of datasets acquired from different sensor systems is necessary to construct a long time-series dataset for remotely sensed land-surface variables. Assessment of the agreement of the data derived from various sources is an important issue in understanding the data continuity through the time-series. Some traditional measures, including correlation coefficient, coefficient of determination, mean absolute error, and root mean square error, are not always optimal for evaluating the data agreement. For this reason, we developed a new agreement coefficient for comparing two different images. The agreement coefficient has the following properties: non-dimensional, bounded, symmetric, and distinguishable between systematic and unsystematic differences. The paper provides examples of agreement analyses for hypothetical data and actual remotely sensed data. The results demonstrate that the agreement coefficient does include the above properties, and therefore is a useful tool for image comparison.

Subfields: Ecology, Media Technology

View Paper: https://www.semanticscholar.org/paper/f65d9a74a461ec27d430990fedf41b0338cb397e

How researchers see this paper

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

In Ecology

In Media Technology

Stephen V. Stehman

Citations in the corpus

Sampling designs for accuracy assessment of land cover (2009, background): “Pontius and Cheuk (2006) and Pontius et al. (2008) suggest alternative parameters to quantify accuracy, and Ji and Gallo (2006) provide a general review of parameters potentially applicable in this setting.”