US 12,272,047 B2
Residue measurement from machine learning based processing of substrate images
Sivakumar Dhandapani, San Jose, CA (US); Arash Alahgholipouromrani, San Jose, CA (US); Dominic J. Benvegnu, La Honda, CA (US); Jun Qian, Sunnyvale, CA (US); and Kiran Lall Shrestha, San Jose, CA (US)
Assigned to Applied Materials, Inc., Santa Clara, CA (US)
Filed by Applied Materials, Inc, Santa Clara, CA (US)
Filed on Oct. 27, 2023, as Appl. No. 18/496,303.
Application 18/496,303 is a continuation of application No. 17/359,307, filed on Jun. 25, 2021, granted, now 11,836,913.
Claims priority of provisional application 63/045,782, filed on Jun. 29, 2020.
Prior Publication US 2024/0054634 A1, Feb. 15, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06T 7/00 (2017.01); B24B 37/013 (2012.01); G06T 7/60 (2017.01)
CPC G06T 7/001 (2013.01) [B24B 37/013 (2013.01); G06T 7/60 (2013.01); G06T 2207/10024 (2013.01); G06T 2207/20021 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30148 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A method of training a neural network, comprising:
obtaining ground truth measurements of amounts of residue on a calibration substrate at a plurality of locations, each location at a defined position for a die being fabricated on the substrate;
obtaining a plurality of color images of the calibration substrate, each color image corresponding to a region for a die being fabricated on the substrate; and
training a neural network to convert color images of die regions from an in-line substrate imager to residue amount measurements for a top layer in each die region of the die regions, the training performed using training data that includes the plurality of color images and ground truth measurements of amounts of residue with each respective color image paired with a ground truth measurement for an amount of residue for each die region associated with the respective color image.