Developing Collaborative Classifiers using an Expert-based Model

Giorgos Mountrakis, Raymond Watts, Lori Luo, Jida Wang
2009 Photogrammetric Engineering and Remote Sensing  
This paper presents a hierarchical, multi-stage adaptive strategy for image classification. We iteratively apply various classification methods (e.g., decision trees, neural networks), identify regions of parametric and geographic space where accuracy is low, and in these regions, test and apply alternate methods repeating the process until the entire image is classified. Currently, classifiers are evaluated through human input using an expert-based system; therefore, this paper acts as the
more » ... f of concept for collaborative classifiers. Because we decompose the problem into smaller, more manageable sub-tasks, our classification exhibits increased flexibility compared to existing methods since classification methods are tailored to the idiosyncrasies of specific regions. A major benefit of our approach is its scalability and collaborative support since selected low-accuracy classifiers can be easily replaced with others without affecting classification accuracy in high accuracy areas. At each stage, we develop spatially explicit accuracy metrics that provide straightforward assessment of results by non-experts and point to areas that need algorithmic improvement or ancillary data. Our approach is demonstrated in the task of detecting impervious surface areas, an important indicator for human-induced alterations to the environment, using a 2001 Landsat scene from Las Vegas, Nevada. and classification error that offers opportunities for synthesis of improved performance through application of different algorithms to different regions of geographic and parameter space. We introduce a strategy for applying such a hybrid process. The central premises of this paper are: (a) that complexity is not high in all parts of the input space, (b) that different algorithms can be applied in different parts, thereby adaptively matching algorithmic complexity to image complexity, and (c) that classification accuracy improvements can be achieved with this approach by establishing a framework for progressive accuracy increase using hybrid classifiers. These premises translate to three simple operating principles that are admittedly cumbersome in application at this early stage of evolution:
doi:10.14358/pers.75.7.831 fatcat:won24zgmcff4vjkkfuuiwlrk4i