Deep learning based defect classification and detection in SEM images: a mask R-CNN approach

Abstract

In this research work, we have demonstrated the application of Mask-RCNN (Regional Convolutional Neural Network), a deep-learning algorithm for computer vision and specifically object detection, to semiconductor defect inspection domain. Stochastic defect detection and classification during semiconductor manufacturing has grown to be a challenging task as we continuously shrink circuit pattern dimensions (e.g., for pitches less than 32 nm). Defect inspection and analysis by state-of-the-art optical and e-beam inspection tools is generally driven by some rule-based techniques, which in turn often causes to misclassification and thereby necessitating human expert intervention. In this work, we have revisited and extended our previous deep learning-based defect classification and detection method towards improved defect instance segmentation in SEM images with precise extent of defect as well as generating a mask for each defect category/instance. This also enables to extract and calibrate each segmented mask and quantify the pixels that make up each mask, which in turn enables us to count each categorical defect instances as well as to calculate the surface area in terms of pixels. We are aiming at detecting and segmenting different types of inter-class stochastic defect patterns such as bridge, break, and line collapse as well as to differentiate accurately between intra-class multi-categorical defect bridge scenarios (as thin/single/multi-line/horizontal/non-horizontal) for aggressive pitches as well as thin resists (High NA applications). Our proposed approach demonstrates its effectiveness both quantitatively and qualitatively.

SEMI-PointRend: improved semiconductor wafer defect classification and segmentation as rendering

“…Recently, Dey et al 5 proposed defect classification and segmentation in SEM images using the Mask R-CNN 6 model architecture. Mask R-CNN is a model that adds a mask header network to Faster R-CNN 8 to enable mask output in addition to classification and bounding boxes.…”

“…Most of these previous works either only classify an image as containing a defect of a certain type or predict the location of a defect in the form of a minimal bounding-box, which we term as defect detection, which generally ignores the precise geometry of the defect pattern itself. Dey et al 5 used the Mask R-CNN 6 neural network architecture to predict the exact pixels in an image which describe a defect pattern precisely. This task is also known as instance segmentation.…”

Addressing misclassification costs in machine learning through asymmetric loss functions

“…The Authors examined this for simulated optical imaging of intentional defect array wafers. 19 Others have focused on SEM images of defects using ML methods such as CNNs [20][21][22][23] and autoencoders. 24…”

Training procedure for scanning electron microscope 3D surface reconstruction using unsupervised domain adaptation with simulated data

“…Deep learning has already been applied to various tasks relevant to SEM images, such as line edge roughness (LER) estimation, 35 denoising, 36,37 and defect inspection. 38 More specifically, similar to the method presented in this work, networks with cyclic losses have also been applied to SEM data. Examples of the latter approach aim at enhancing the quality of the SEM images 39 or mapping the chip layout designs to SEM images.…”