2008 · 10 citations · 14 references
EngineeringManufacturing EngineeringComputer VisionDigital ImagingMachine VisionArtifact RecognitionPrinting TechniquesDigital Image EnhancementHistogramsPrintabilityPrintability EnhancementSurface DefectsSurface InspectionThresholdingVisual Inspection Systems
Fine and sparse details, as defined in this study, occur in several applications where image processing can be applied. The definition of fine and sparse details is given in the domain of gray level histograms by establishing sufficient statistical properties. In experiments, artificial data was generated from a model based on the definition, and several thresholding methods were compared. Based on the experiments on artificial data, it seems that the minimum error thresholding by Kittler and Illingworth outperforms the other methods. The result was verified using images from IGT picking paper printability assessment where small surface defects must be detected. Furthermore, image enhancement and a minor adjustment to the minimum error thresholding are introduced to achieve optimal performance in IGT picking paper printability assessment. 1
14