IP Library › Granted Patent US 12,614,256
Granted Patent B2
US 12,614,256 · App. 18/128,184 · Granted Apr 28, 2026

Shot noise reduction using frame averaging

Inventors: Abdurrahman Sezginer (Monte Sereno, CA); Patrick McBride (Davis, CA); Indrasen Bhattacharya (Milpitas, CA); Robert M. Danen (San Jose, CA)
Assignee: KLA Corporation
G06T5/70G06T5/20G06T7/001G06V20/50G06T2207/20224G06T2207/30148
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,614,256
App. No.
18/128,184
Granted
Apr 28, 2026
Kind
B2
Abstract

An inspection system includes a controller including one or more processors configured to execute program instructions. The program instructions cause the one or more processors to receive at least a first portion of a first set of repeat swaths of a first scan path of a sample. The program instructions cause the one or more processors to generate an image by averaging the first portion of the first set of repeat swaths. Averaging the first portion of the first set of repeat swaths reduces a noise in the image. The program instructions cause the one or more processors to detect one or more defects in an inspection region of the sample using the image.

Claims (61)

1 . An inspection system comprising:

a controller including one or more processors configured to execute program instructions causing the one or more processors to:

receive at least a first portion of a first set of repeat swaths of a first scan path of a sample;

align and scale the first portion of the first set of repeat swaths resulting in a common number of pixels and a common pixel location for each pixel of the first portion of the first set of repeat swaths;

generate an image by averaging the first portion of the first set of repeat swaths; wherein averaging the first portion of the first set of repeat swaths reduces a noise in the image,

wherein aligning the first portion of the first set of repeat swaths occurs before averaging the first portion; and

detect one or more defects in an inspection region of the sample using the image.

2 . The inspection system of claim 1 , wherein one of:

the image is an average inspection image of the inspection region; or

the image is a reference image.

3 . The inspection system of claim 2 , wherein the image is the average inspection image; wherein the program instructions cause the one or more processors to:

receive a plurality of inspection images; wherein the plurality of inspection images are repeated images of the inspection region; wherein the plurality of inspection images includes a first plurality of signal-to-noise ratios; wherein the plurality of inspection images each include the noise contributing to the first plurality of signal-to-noise ratios; and

average the plurality of inspection images to generate the average inspection image; wherein the average inspection image includes a second signal-to-noise ratio; wherein the second signal-to-noise ratio is greater than each of the first plurality of signal-to-noise ratios.

4 . The inspection system of claim 3 , wherein the inspection system generates the plurality of inspection images in a scanning mode from the first portion of the first set of repeat swaths.

5 . The inspection system of claim 2 , wherein the program instructions cause the one or more processors to subtract the reference image from the average inspection image to generate a difference image.

6 . The inspection system of claim 5 , wherein the program instructions cause the one or more processors to detect one or more defects in the inspection region based on the difference image.

7 . The inspection system of claim 5 , wherein the program instructions cause the one or more processors to:

apply a difference filter to the difference image to generate a filtered difference image; and

detect one or more defects in the inspection region based on the filtered difference image.

8 . The inspection system of claim 5 , wherein the program instructions cause the one or more processors to classify the one or more defects as at least one of a nuisance or a defect of interest.

9 . The inspection system of claim 1 , wherein the program instructions cause the one or more processors to align the first portion using sync interpolation.

10 . The inspection system of claim 2 , comprising one or more optical inspection sub-systems; wherein the one or more optical inspection sub-systems generate the first set of repeat swaths of the first scan path; wherein the controller is communicatively coupled to the one or more optical inspection sub-systems.

11 . The inspection system of claim 10 , comprising a stage; wherein the stage translates causing the one or more optical inspection sub-systems to generate the first set of repeat swaths of the first scan path.

12 . The inspection system of claim 11 , comprising one or more of an encoder or an interferometer; wherein one or more of the encoder or the interferometer are configured to monitor the translation of the stage; wherein the program instructions cause the one or more processors to control the translation of the stage to align the first set of repeat swaths of the first scan path.

13 . The inspection system of claim 10 , wherein the one or more processors receive at least the first portion of the first set of repeat swaths of the first scan path from the one or more optical inspection sub-systems.

14 . The inspection system of claim 10 , wherein the one or more optical inspection sub-systems comprise two or more optical inspection sub-systems.

15 . The inspection system of claim 10 , wherein the one or more optical inspection sub-systems comprise a single optical inspection sub-system.

16 . The inspection system of claim 10 , wherein the one or more optical inspection sub-systems comprises at least a first optical mode; wherein the one or more optical inspection sub-systems generates the first set of repeat swaths using the first optical mode.

17 . The inspection system of claim 16 , wherein the first optical mode comprises a first illumination wavelength, a first illumination polarization, and a first illumination angle for the one or more optical inspection sub-systems.

18 . The inspection system of claim 2 , wherein the image is the average inspection image of the inspection region; wherein the image is the reference image in a subsequent inspection.

19 . The inspection system of claim 1 , wherein the first set of repeat swaths comprises at least two repeat swaths.

20 . The inspection system of claim 19 , wherein the first set of repeat swaths comprises between two and four repeat swaths.

21 . The inspection system of claim 1 , wherein the first set of repeat swaths comprises at least four repeat swaths.

22 . The inspection system of claim 1 , wherein the image is one image of a plurality of images used by the processors to generate a reference image; wherein the program instructions cause the one or more processors to generate the reference image from the plurality of images; wherein the one image is generated by scan averaging.

23 . The inspection system of claim 22 , wherein the inspection region includes one or more patterned features; wherein the plurality of images are from a plurality of reference regions each comprising the one or more patterned features.

24 . The inspection system of claim 1 , wherein the program instructions cause the one or more processors to receive a second portion of a second set of repeat swaths of a second scan path of the sample and average the second portion of the second set of repeat swaths; wherein the image comprises at least the first portion and the second portion.

25 . The inspection system of claim 1 , wherein the noise comprises at least one of a readout noise or a shot noise.

26 . An inspection system comprising:

one or more optical inspection sub-systems;

a controller communicatively coupled to the one or more optical inspection sub-systems, the controller including one or more processors to execute program instructions causing the one or more processors to:

receive at least a first portion of a first set of repeat swaths of a first scan path of a sample;

align and scale the first portion of the first set of repeat swaths resulting in a common number of pixels and a common pixel location for each pixel of the first portion of the first set of repeat swaths;

generate an image by averaging the first portion of the first set of repeat swaths; wherein averaging the first portion of the first set of repeat swaths reduces a noise in the image,

wherein aligning the first portion of the first set of repeat swaths occurs before averaging the first portion; and

detect one or more defects in an inspection region of the sample using the image.

27 . The inspection system of claim 26 , wherein at least one of:

the image is an average inspection image of the inspection region;

the image is a reference image; or

the image is one image of a plurality of images used by the processors to generate the reference image.

28 . The inspection system of claim 27 , wherein the image is the average inspection image; wherein the program instructions cause the one or more processors to:

receive a plurality of inspection images; wherein the plurality of inspection images are repeated images of the inspection region; wherein the plurality of inspection images includes a first plurality of signal-to-noise ratios; wherein the plurality of inspection images each include the noise contributing to the first plurality of signal-to-noise ratios; and

average the plurality of inspection images to generate the average inspection image; wherein the average inspection image includes a second signal-to-noise ratio; wherein the second signal-to-noise ratio is greater than each of the first plurality of signal-to-noise ratios.

29 . The inspection system of claim 27 , wherein the image is the one image of the plurality of images; wherein the program instructions cause the one or more processors to generate the reference image from the plurality of images.

30 . The inspection system of claim 29 , wherein the inspection region includes one or more patterned features; wherein the plurality of images are from a plurality of reference regions each comprising the one or more patterned features.

31 . A method comprising:

receiving at least a first portion of a first set of repeat swaths of a first scan path of a sample;

aligning and scaling the first portion of the first set of repeat swaths resulting in a common number of pixels and a common pixel location for each pixel of the first portion of the first set of repeat swaths;

generating an image by averaging the first portion of the first set of repeat swaths; wherein averaging the first portion of the first set of repeat swaths reduces a noise in the image,

wherein aligning the first portion of the first set of repeat swaths occurs before averaging the first portion; and

detecting one or more defects in an inspection region of the sample using the image.

32 . The inspection system of claim 1 , wherein the first portion of the first set of repeat swaths is pre-aligned before receiving the first portion of the first set of repeat swaths of the first scan path of the sample.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 16, 2023
From: SEZGINER, ABDURRAHMAN; MCBRIDE, PATRICK; BHATTACHARYA, INDRASEN; DANEN, ROBERT M.
To: KLA CORPORATION
Reel/Frame 063974/0914 →
Continuity (2)
Provisional Application 63326268 · Mar 31, 2022
Related Publication 20230351553A1 · Nov 2, 2023
References Cited (81)
US 6987892B2 · Edgar · 2006 [cited by applicant]
US 7877722B2 · Duffy et al. · 2011 [cited by applicant]
US 8126255B2 · Bhaskar et al. · 2012 [cited by applicant]
US 8611639B2 · Kulkarni et al. · 2013 [cited by applicant]
US 9183624B2 · Karsenti et al. · 2015 [cited by applicant]
US 9262821B2 · Shifrin et al. · 2016 [cited by applicant]
US 9518934B2 · Chen et al. · 2016 [cited by applicant]
US 9613411B2 · Konuru et al. · 2017 [cited by applicant]
US 9619876B2 · Huang et al. · 2017 [cited by applicant]
US 10346740B2 · Zhang et al. · 2019 [cited by applicant]
US 10395358B2 · Brauer et al. · 2019 [cited by applicant]
US 10440299B2 · Lin et al. · 2019 [cited by applicant]
US 10504213B2 · Brauer · 2019 [cited by applicant]
US 10533848B2 · Shchegrov et al. · 2020 [cited by applicant]
US 10535131B2 · Maher et al. · 2020 [cited by applicant]
US 10557802B2 · Brauer et al. · 2020 [cited by applicant]
US 10599944B2 · Lee et al. · 2020 [cited by applicant]
US 10801968B2 · Brauer · 2020 [cited by applicant]
US 10923317B2 · Huang et al. · 2021 [cited by applicant]
US 10943838B2 · Hoo et al. · 2021 [cited by applicant]
US 11010885B2 · Brauer et al. · 2021 [cited by applicant]
US 11120546B2 · Brauer et al. · 2021 [cited by applicant]
US 11270430B2 · Sezginer et al. · 2022 [cited by applicant]
US 11415526B2 · Duffy et al. · 2022 [cited by applicant]
US 11415531B2 · Gaind et al. · 2022 [cited by applicant]
US 11416982B2 · Brauer et al. · 2022 [cited by applicant]
US 11494924B2 · Noormofidi et al. · 2022 [cited by applicant]
US 11550309B2 · Soltanmohammadi et al. · 2023 [cited by applicant]
US 11551348B2 · Zhang et al. · 2023 [cited by applicant]
US 20030012442A1 · Hashimoto · 2003 [cited by examiner]
US 20040196453A1 · Some · 2004 [cited by applicant]
US 20080165345A1 · Hill · 2008 [cited by applicant]
US 20080291454A1 · Cai et al. · 2008 [cited by applicant]
US 20080317358A1 · Bressan et al. · 2008 [cited by applicant]
US 20090080759A1 · Bhaskar et al. · 2009 [cited by applicant]
US 20100329540A1 · Bhaskar et al. · 2010 [cited by applicant]
US 20110314049A1 · Poirier et al. · 2011 [cited by applicant]
US 20120308112A1 · Hu et al. · 2012 [cited by applicant]
US 20150006103A1 · Wel et al. · 2015 [cited by applicant]
US 20170116727A1 · Kim et al. · 2017 [cited by applicant]
US 20180082415A1 · Sezginer et al. · 2018 [cited by applicant]
US 20180144442A1 · Brauer · 2018 [cited by applicant]
US 20190033728A1 · Takagi et al. · 2019 [cited by applicant]
US 20190122913A1 · Lauber et al. · 2019 [cited by applicant]
US 20190130551A1 · Chen · 2019 [cited by examiner]
US 20190287232A1 · Brauer · 2019 [cited by applicant]
US 20200025689A1 · Gaind et al. · 2020 [cited by applicant]
US 20200193588A1 · Brauer et al. · 2020 [cited by applicant]
US 20210010945A1 · Brauer et al. · 2021 [cited by applicant]
US 20210109041A1 · Gaind et al. · 2021 [cited by applicant]
US 20210201460A1 · Gong et al. · 2021 [cited by applicant]
US 20210366103A1 · Zhang et al. · 2021 [cited by applicant]
US 20210381989A1 · Hiroi et al. · 2021 [cited by applicant]
US 20220050061A1 · Bar · 2022 [cited by applicant]
JP 2004327950A · 2004 [cited by applicant]
JP 4789630B2 · 2011 [cited by applicant]
JP 2015115504A · 2015 [cited by examiner]
KR 1020080028278A · 2008 [cited by applicant]
KR 1020170100710A · 2017 [cited by applicant]
TW 202041850A · 2020 [cited by applicant]
TW 202130994A · 2021 [cited by applicant]
WO 2013134068A1 · 2013 [cited by applicant]
WO 2019167151A1 · 2019 [cited by applicant]
WO 2020035285A1 · 2020 [cited by applicant]
Korean Intellectual Property Office, International Search Report and Written Opinion for International Application No. PCT/US2023/017004, Jul. 12, 2023, 9 pages. [cited by applicant]
Korean Intellectual Property Office, International Search Report and Written Opinion for International Application No. PCT/US2023/017007, Jul. 11, 2023, 7 pages. [cited by applicant]
Korean Intellectual Property Office, International Search Report and Written Opinion for International Application No. PCT/US2023/017008, Jul. 31, 2023, 6 pages. [cited by applicant]
Korean Intellectual Property Office, International Search Report and Written Opinion for International Application No. PCT/US2023/017001, Jul. 24, 2023, 9 pages. [cited by applicant]
Hiroi et al. (“Robust Defect Detection System Using Double Reference Image Averaging for High Throughout SEM Inspection Tool,” The 17th Annual SEMI/IEEE ASMC 2006 Conference; Date of Conference: May 22-24, 2006) (Year: … [cited by applicant]
Kim et al. (“Improved Non-parametric Subtraction for Detection of Wafer Defect,” 5th International Symposium on Image and Signal Processing and Analysis; Date of Conference: Sep. 27-29, 2007) (Year: 2007). [cited by applicant]
Amano, “Defect Detection starter (U-Net), Severstal: Steel Defect Detection”, Aug. 3, 2019, 25 pages. [cited by applicant]
Kälber et al., “U-Net based Zero-hour Defect Inspection of Electronic Components andSemiconductors,” VISAPP 2021—16th International Conference on Computer Vision Theory and Applications, 2021, 9 pages, SCITEPRESS—Scienc… [cited by applicant]
Khanmohammadi et al., “A Gaussian mixture model based discretization algorithm for associative classification of medical data,” Expert Systems With Applications, Apr. 8, 2016, 11 pages, Elsevier Ltd. [cited by applicant]
Ren et al., “A Generic Deep-Learning-Based Approach for Automated Surface Inspection,” IEEE Transactions on Cybernetics, 12 pages, vol. 48, No. 3, Mar. 2018, IEEE. [cited by applicant]
Ronneberger et al., “U-Net: Convolutional Networks for Biomedical Image Segmentation,” May 18, 2015, 8 pages, Computer Science Department and BIOSS Centre for Biological Signalling Studies, University of Freiburg, Germa… [cited by applicant]
Salimans et al., “PIXELCNN++: Improving the Pixelcnn With Discretized Logistic Mixture Likelihood and Other Modifications,” Conference Paper at ICLR 2017, Jan. 19, 2017, 10 pages. [cited by applicant]
Van Den Oord et al., “Conditional Image Generation with PixelCNN Decoders,” Jun. 18, 2016, 13 pages. [cited by applicant]
Yamashita et al., “Convolutional neural networks: an overview and application in radiology,” Insights into Imaging, Jun. 22, 2018, 19 pages, Springer Nature. [cited by applicant]
Yun et al., “Automated defect inspection system for metal surfaces based on deep learning and data augmentation,” Journal of Manufacturing Systems, May 15, 2020, 8 pages, No. 55, Elsevier Ltd. [cited by applicant]
Zheng et al., “Recent advances in surface defect inspection of industrial products using deep learning techniques,” Survey Paper, The International Journal of Advanced Manufacturing Technology, Jan. 25, 2021, 24 pages, … [cited by applicant]
Taiwan Patent Office, Office Action for Taiwan Application No. 112112378 dated Jan. 6, 2026, 22 pages (with translation). [cited by applicant]