Segmentation d'images de semiconducteur pour la détection de défauts

This paper introduces a segmentation algorithm s uitable for semiconductor wafer images generated by optical inspection tools. The primary application of this work is content-bas ed region segmentation for automatic threshold sele ction during recipe generation in dieto-die wafer inspection. Structures associated with d fferent functional areas lead to different level s of noise in the difference image during the defect detection process. The ability to automa ically create a mask to separate the different str uctures and materials is necessary to determine local thresholds for each area and thus t o improve the signal-to-noise ratio. A supervised s gmentation based on the discrete wavelet transform is used to segment a whole die to create a mask. During the inspection, the mask is applied on the difference image, and the threshold is automatically set as a function of the noise within the region and the thresholding c oefficient specific to that region. The use of the segmented region in content-based threshold defect detection improves the number of defects det ected, and reduces the number of false detections.