A wafer map contains a graphical representation of the locations about defect pattern on the semiconductor wafer, which can provide useful information for quality engineers. Various defect patterns ...occur due to increasing wafer sizes and decreasing features sizes, which makes it very complex and unreliable process to identify them. In this paper, we propose a voting ensemble classifier with multi-types features to identify wafer map defect patterns in semiconductor manufacturing. Our research contents can be summarized as follows. First, three distinctive features such as density-, geometry-, and radon-based features were extracted from raw wafer images. Then, we applied four machine learning classifiers, namely logistic regression (LR), random forests (RFs), gradient boosting machine (GBM), and artificial neural network (ANN), and trained them using extracted features of original data set. Then their results were combined with a soft voting ensemble (SVE) technique which assigns higher weights to the classifiers with respect to their prediction accuracy. Consequently, we got performance measures with accuracy, precision, recall, <inline-formula> <tex-math notation="LaTeX">{F} </tex-math></inline-formula>-measure, and AUC score of 95.8616%, 96.9326%, 96.9326%, 96.7124%, and 99.9114%, respectively. These results show that the SVE classifier with proposed multi-types features outperformed regular machine learning-based classifiers for wafer maps defect detection.
Wafer maps contain information about various defect patterns on the wafer surface and automatic classification of these defects plays a vital role to find their root causes. Semiconductor engineers ...apply various methods for wafer defect classification such as manual visual inspection or machine learning-based algorithms by manually extracting useful features. However, these methods are unreliable, and their classification performance is also poor. Therefore, this paper proposes a deep learning-based convolutional neural network for automatic wafer defect identification (CNN-WDI). We applied a data augmentation technique to overcome the class-imbalance issue. The proposed model uses convolution layers to extract valuable features instead of manual feature extraction. Moreover, state-of-the-art regularization methods such as batch normalization and spatial dropout are used to improve the classification performance of the CNN-WDI model. The experimental results comparison using a real wafer dataset shows that our model outperformed all previously proposed machine learning-based wafer defect classification models. The average classification accuracy of the CNN-WDI model with nine different wafer map defects is 96.2%, which is an increment of 6.4% from the last highest average accuracy using the same dataset.
Infrared (IR)‐to‐visible up‐conversion device allows a low‐cost, pixel‐free IR imaging over the conventional expensive compound semiconductor‐based IR image sensors. However, the external quantum ...efficiency has been low due to the integration of an IR photodetector and a light‐emitting diode (LED). Herein, by inducing a strong micro‐cavity effect, a highly efficient top‐emitting IR‐to‐visible up‐conversion device is demonstrated where PbS quantum dots IR‐absorbing layer is integrated with a phosphorescent organic LED. By optimizing the optical cavity length between indium tin oxide (ITO)/thin Ag/ITO anode and semi‐transparent Mg:Ag top cathode, the up‐conversion device yields 15.7% of photon‐to‐photon conversion efficiency from the top‐emission. The high efficiency can be achieved under a low IR transmission through the semi‐reflective anode. Finally, pixel‐free IR imaging is demonstrated using the up‐conversion device, boosting the effect of micro‐cavity on the brightness and the contrast of an IR image.
High efficiency top‐emitting infrared (IR)‐to‐visible up‐conversion device is demonstrated by exploiting microcavity effect. Compared to up‐conversion device using conventional indium tin oxide (ITO) electrode, ITO/Ag/ITO reflective electrode offers a strong optical resonance toward the top side, yielding 15.7 % IR‐to‐photon conversion efficiency. Using the microcavity effect, pixel‐free IR imaging is demonstrated with higher brightness and image contrast.
This study aimed to develop a high-performance deep learning algorithm to differentiate Stafne’s bone cavity (SBC) from cysts and tumors of the jaw based on images acquired from various panoramic ...radiographic systems. Data sets included 176 Stafne’s bone cavities and 282 odontogenic cysts and tumors of the mandible (98 dentigerous cysts, 91 odontogenic keratocysts, and 93 ameloblastomas) that required surgical removal. Panoramic radiographs were obtained using three different imaging systems. The trained model showed 99.25% accuracy, 98.08% sensitivity, and 100% specificity for SBC classification and resulted in one misclassified SBC case. The algorithm was approved to recognize the typical imaging features of SBC in panoramic radiography regardless of the imaging system when traced back with Grad-Cam and Guided Grad-Cam methods. The deep learning model for SBC differentiating from odontogenic cysts and tumors showed high performance with images obtained from multiple panoramic systems. The present algorithm is expected to be a useful tool for clinicians, as it diagnoses SBCs in panoramic radiography to prevent unnecessary examinations for patients. Additionally, it would provide support for clinicians to determine further examinations or referrals to surgeons for cases where even experts are unsure of diagnosis using panoramic radiography alone.
Abstract Objective Stent graft-induced new entry (SINE) has been increasingly observed after thoracic endovascular aortic repair (TEVAR) for Stanford type B aortic dissection. SINE is often life ...threatening, and reintervention is required. This study investigated risk factors for SINE after TEVAR. Methods From July 2001 to June 2013, we retrospectively analyzed data from 79 patients who underwent TEVAR for Stanford type B aortic dissection. TEVAR was performed in 17 patients ≤2 weeks (acute) after the diagnosis of aortic dissection and in the remaining 62 patients >2 weeks (chronic) after diagnosis. Forty-two of the patients underwent TEVAR with modified stent graft with an “inwardly bent” margin, and the remaining 37 underwent TEVAR with a conventional stent graft. The maximal diameter, minimal diameter, mean diameter, circumference, and area of the true lumen were analyzed. Taper ratio and oversizing ratio were evaluated and compared between the SINE and non-SINE groups, and cutoff values of taper ratio and oversizing ratio for prediction of SINE were determined using receiver-operating characteristic curve analysis. The cumulative incidence of SINE was estimated with the Kaplan-Meier method. The multivariate Cox proportional hazards model was used to identify independent predictive variables for SINE. Results SINE occurred in 21 patients (26.5%) and occurred more frequently in patients with chronic dissection than in those with acute dissection (32.3% vs 5.9%; P = .032). The Kaplan-Meier curves were significantly different ( P = .016) between these groups. The incidence of SINE events was not significantly different between the modified stent group and nonmodified stent group (23.8% vs 36.0%; P = .284). The taper ratio and oversizing ratio by maximal diameter, mean diameter, circumference, and area were significantly higher in the SINE group than in the non-SINE group, and Kaplan-Meier curves were significantly different between groups above and below optimal cutoff value ( P < .0005 to .003). According to multivariate analysis, the hazard ratios of chronic aortic dissection were 6.30 (95% confidence interval, 0.83-47.74; P = .075) to 7.80 (95% confidence interval, 1.03-59.07; P = .047). The taper ratio and oversizing ratio calculated by maximal diameter, mean diameter, circumference, and area were independent predictors of the development of SINE. Conclusions Distal oversizing of the stent graft was an independent predictor of the development of SINE. Appropriate size selection of stent graft without distal oversizing might reduce the risk of late SINE events.
Traditional regression-based approaches do not provide good results in diagnosis and prediction of occurrences of cardiovascular diseases (CVD). Therefore, the goal of this paper is to propose a deep ...learning–based prediction model of occurrence of major adverse cardiac events (MACE) during the 1, 6, 12 month follow-up after hospital admission in acute myocardial infarction (AMI) patients using knowledge mining. We used the Korea Acute Myocardial Infarction Registry (KAMIR) dataset, a cardiovascular disease database registered in 52 hospitals in Korea between 1 January, 2005, and 31 December, 2008. Among 14,885 AMI patients, 10,813 subjects in age from 20 to 100 years with the 1-year follow-up traceability without coding errors were finally selected. For our experiment, the training/validation/test dataset split is 60/20/20 by random sampling without replacement. The preliminary deep learning model was first built by applying training and validation datasets and then a new preliminary deep learning model was generated using the best hyperparameters obtained from random hyperparameter grid search. Lastly, the preliminary prediction model of MACE occurrences in AMI patients is evaluated by test dataset. Compared with conventional regression-based models, the performances of machine/deep learning–based prediction models of the MACE occurrence in patients with AMI, including deep neural network (DNN), gradient boosting machine (GBM), and generalized linear model (GLM), are also evaluated through a matrix with sensitivity, specificity, overall accuracy, and the area under the ROC curve (AUC). The prediction results of the MACE occurrence during the 1, 6, and 12-month follow-up in AMI patients were the AUC of DNN (1 M 0.97, 6 M 0.94, 12 M 0.96), GBM (0.96, 0.95, 0.96), and GLM (0.76, 0.67, 0.72) in machine learning–based models as well as GRACE (0.75, 0.72, 0.76) in regression model. Compared with previous models, our deep learning–based prediction models significantly had the accuracy of 95% or higher and outperformed all machine learning and regression-based prediction models. This paper was the first trial of deep learning–based prediction model of the MACE occurrence in AMI clinical data. We found that the proposed prediction model applied different risk factors except the attribute “age” by using knowledge mining and directly used the raw data as input.
Objective Some researchers have studied about early prediction and diagnosis of major adverse cardiovascular events (MACE), but their accuracies were not high. Therefore, this paper proposes a soft ...voting ensemble classifier (SVE) using machine learning (ML) algorithms. Methods We used the Korea Acute Myocardial Infarction Registry dataset and selected 11,189 subjects among 13,104 with the 2-year follow-up. It was subdivided into two groups (ST-segment elevation myocardial infarction (STEMI), non ST-segment elevation myocardial infarction NSTEMI), and then subdivided into training (70%) and test dataset (30%). Third, we selected the ranges of hyper-parameters to find the best prediction model from random forest (RF), extra tree (ET), gradient boosting machine (GBM), and SVE. We generated each ML-based model with the best hyper-parameters, evaluated by 5-fold stratified cross-validation, and then verified by test dataset. Lastly, we compared the performance in the area under the ROC curve (AUC), accuracy, precision, recall, and F-score. Results The accuracies for RF, ET, GBM, and SVE were (88.85%, 88.94%, 87.84%, 90.93%) for complete dataset, (84.81%, 85.00%, 83.70%, 89.07%) STEMI, (88.81%, 88.05%, 91.23%, 91.38%) NSTEMI. The AUC values in RF were (98.96%, 98.15%, 98.81%), ET (99.54%, 99.02%, 99.00%), GBM (98.92%, 99.33%, 99.41%), and SVE (99.61%, 99.49%, 99.42%) for complete dataset, STEMI, and NSTEMI, respectively. Consequently, the accuracy and AUC in SVE outperformed other ML models. Conclusions The performance of our SVE was significantly higher than other machine learning models (RF, ET, GBM) and its major prognostic factors were different. This paper will lead to the development of early risk prediction and diagnosis tool of MACE in ACS patients.
Here, we report gold nanoparticle-coated starch magnetic beads (AuNP@SMBs) that were prepared by in situ synthesis of AuNPs on the surface of SMBs. Upon functionalization of the surface with a ...specific antibody, the immuno-AuNP@SMBs were found to be effective in separating and concentrating the target pathogenic bacteria, Escherichia coli O157:H7, from an aqueous sample as well as providing a hotspot for surface-enhanced Raman scattering (SERS)-based detection. We employed a bifunctional linker protein, 4× gold-binding peptide-tagged Streptococcal protein G (4GS), to immobilize antibodies on AuNP@SMBs and AuNPs in an oriented form. The linker protein also served as a Raman reporter, exhibiting a strong and unique fingerprint signal during the SERS measurement. The amplitude of the SERS signal was shown to have a good correlation with the concentration of target bacteria ranging from 100 to 105 CFU/mL. The detection limit was determined to be as low as a single cell, and the background signals derived from nontarget bacteria were negligible due to the excellent specificity and colloidal stability of the immuno-AuNP@SMBs and SERS tags. The highly sensitive nature of the SERS-based detection system will provide a promising means to detect the pathogenic microorganisms in food or clinical specimen.
Wafer maps contain information about defects and clustered defects that form failure patterns. Failure patterns exhibit the information related to defect generation mechanisms. The accurate ...classification of failure patterns in wafer maps can provide crucial information for engineers to recognize the causes of the fabrication problems. In this paper, we proposed a decision tree ensemble learning-based wafer map failure pattern recognition method based on radon transform-based features. Radon transform is applied on raw wafer map data to generate the new features which are exhibiting the geometric information of failure patterns in wafer map. Decision tree algorithm is applied to build decision tree ensemble and the final decision is made by aggregating the prediction results of decision trees. The effectiveness of the proposed method has been verified by using the real world wafer map data set (WM-811K).
PDF
