THIS FEATURE CLASS CONTAINS PRELIMINARY STRUCTURE (BUILDING) POLYGONS FOR THE STATE OF Mississippi. The structures (buildings) are represented by polygon geometries with an attribute table described elsewhere in the metadata. WorldView-02 and WorldView-03 high resolution multispectral imagery were obtained for Mississippi. The raw imagery (1B product) for Mississippi was then pan-sharpened and orthorectified. Complete coverage of Mississippi with minimal cloud cover includes imagery acquired between 2/06/2011 and 6/5/2018 and spatial resolutions between 0.46-0.85 meters. Once the imagery pre-processing was complete, training data was acquired and multiple convolutional neural network (CNN) models were generated. High Performance Computing (HPC) environments at ORNL were then used to apply the CNN to the imagery resulting in this preliminary structure (building) feature class. Minimal review of this dataset has been completed. Verification and validation of this dataset was conducted by algorithmic model. See ‘Automated’ domain value for VAL_METHOD for more details. Only structure polygon results whose area is greater than 450 square feet are included in this feature class. The attribute schema associated with this feature class has been developed under previous prototype efforts for structure detection and continues to evolve. Population of some attributes is incomplete and identified in the Entity and Attribute Information section of this metadata. The visual representation of the features in this feature class have been improved.