1. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Fragmentation-based hotspot signature extraction. (a) Layout patterns and the Hanan grids shown in dashed lines. (b) Fragmentation-based context characterization within the effective radius. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

2. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Examples of hotspot patterns marked in red. (a) Hotspot resulted from one-dimensional (1-D) patterns only. (b) Hotspot resulted from complex 1-D and 2-D patterns. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

3. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Hotspot model calibration flow. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

4. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Pattern types. (a) 1-D pattern including one long feature; (b) 1-D pattern including two parallel long features; (c) 2-D pattern including one long feature; (d) 2-D mountain-shaped pattern. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

5. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. K-means clustering with k=5. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

6. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. SVM-based data learning flow. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

7. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Layout scanning. (a) Enlarged checking area. (b) Slashes show the overlapped area between adjacent checking areas. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

8. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Full layout hotspot detection flow. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

9. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Pattern examples of detected hotspots. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

10. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Pattern examples of false alarms. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

11. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Pattern examples of nondetected hotspots. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003

12. Date of download: 7/2/2016 Copyright © 2016 SPIE. All rights reserved. Results comparison with different sampling rates for Array_benchmark5. Figure Legend: From: Accurate lithography hotspot detection based on principal component analysis- support vector machine classifier with hierarchical data clustering J. Micro/Nanolith. MEMS MOEMS. 2014;14(1):011003. doi:10.1117/1.JMM.14.1.011003