• To automatically evaluate scenic beauty quality of forestry images
• Important to the USFS for management of timber production and forest growth
  
  
• Extracted diversified features from the forestry image database
  
• Applied various pattern recognition algorithms to these features to classify images into HSBE, MSBE or LSBE
  
• Best error classification rates: 32.2% with SVM, 33.4% with ICA, and 37.7% with PCA