Kla Tencor Corporation

Midea Group Co., Ltd.

Thanh Huy Ha

Ajay Gupta

Zhicai Ou

Yunke Tian

Yi Heng Chen

Brian Duffy

Hedong Yang

Jianwei Wang

Christopher Sears

Scott Allen Young

Mohan Mahadevan

Michael John Van Riet

Chien-Huei Adam Chen

Christopher Maher

Olivier Moreau

Huina Xu

Mohamad Al Jazaery

Yuxiang Gao

Ichiro Honjo

Kumar Raja

Sathyanarayanan Muthusamy

Adam Chien-Huei Chen

Michael Van Riet

A method and system for capturing, by a camera a sequence of frames at respective locations within a portion of an environment; capturing, by an inertial measurement unit, a sequence of inertial odometry data corresponding to the sequence of frames at the respective locations; storing in a queue a data record includes information extracted from processing the respective frame and information from the inertial measurement unit; in accordance with a determination that the sequence of inertial odometry data satisfies a first criterion: calculating a first relative pose between the first frame and the second frame; and in accordance with a determination that a difference between the first relative pose and the information extracted from processing the respective frame satisfy a first threshold: generating an initial map of the portion of the environment based on the first data record and the second data record.

Fast and robust initialization method for feature-based monocular visual SLAM using inertial odometry assistance

System and method for reducing false negatives in object detection, including: extracting an object of interest from a respective image in a first set of training data that includes in distribution (ID) data. For each of the object of interest extracted from the respective image in the first set of training data: fusing the object of interest with an image from a second set of data that does not include any objects of interest to form a fused image; adding the fused image to the training data; and using the training data to train a detection model for object detection.

Method and system for reducing false positives in object detection neural networks caused by novel objects

System and method for providing visual-aided placement recommendation includes obtaining images of a rack configured to hold objects inside a chamber, placement of the plurality of objects on the rack being subject to preset constraints corresponding to characteristics of respective objects of the plurality of objects relative to physical parameters of respective locations on the rack; analyzing the images to determine whether the preset constraints have been violated by placement of objects on the rack; and in accordance with a determination that at least one preset constraint has been violated, generating a first output providing a guidance on proper placement of the first object on the rack that complies with the one or more preset constraints, in accordance with the physical characteristics of the first object relative to the physical parameters of the respective locations on the rack, taking into account of other objects already placed on the rack.

System and method for recommending object placement

System and method for controlling a machine, including: receiving a first image processing model trained to classify an input image into a first class for images containing at least one object of a first type or a second class for images not containing an object of the first type; identifying a subset of inference results that are false positive results; generating a set of new training data from the first set of images, including augmenting an image in the first set of images to obtain a respective plurality of images and labeling the respective plurality of images as containing at least one object of a pseudo first class; training a second image processing model to classify an input image into the first class, the second class, and the first pseudo class; and modifying a device setting of a machine based on an inference result of the second image processing model.

Adjusting machine settings through multi-pass training of object detection models

A method includes: capturing one or more images of an unorganized collection of items inside a first machine; determining one or more item types of the unorganized collection of items from the one or more images, comprising: dividing a respective image in the one or more images into a respective plurality of sub-regions; performing feature detection on the respective plurality of sub-regions to obtain a respective plurality of regional feature vectors, wherein a regional feature vector for a sub-region indicates characteristics for a plurality of predefined local item features for the sub-region; generating an integrated feature vector by combining the respective plurality of regional feature vectors; and applying a plurality of binary classifiers to the integrated feature vector; and selecting a machine setting for the first machine based on the determined one or more clothes type in the unorganized collection of items.

Method and system for controlling machines based on object recognition

Objective lens alignment of a scanning electron microscope review tool with fewer image acquisitions can be obtained using the disclosed techniques and systems. Two different X-Y voltage pairs for the scanning electron microscope can be determined based on images. A second image based on the first X-Y voltage pair can be used to determine a second X-Y voltage pair. The X-Y voltage pairs can be applied at the Q4 lens or other optical components of the scanning electron microscope.

Scanning electron microscope objective lens calibration using X-Y voltages iteratively determined from images obtained using said voltages

Methods and systems for aligning images for a specimen acquired with different modalities are provided. One method includes acquiring information for a specimen that includes at least first and second images for the specimen. The first image is acquired with a first modality different than a second modality used to acquire the second image. The method also includes inputting the information into a learning based model. The learning based model is included in one or more components executed by one or more computer systems. The learning based model is configured for transforming one or more of the at least first and second images to thereby render the at least the first and second images into a common space. In addition, the method includes aligning the at least the first and second images using results of the transforming. The method may also include generating an alignment metric using a classifier.

Learning based approach for aligning images acquired with different modalities

Methods and systems for determining parameter(s) of a metrology process to be performed on a specimen are provided. One system includes one or more computer subsystems configured for automatically generating regions of interest (ROIs) to be measured during a metrology process performed for the specimen with the measurement subsystem based on a design for the specimen. The computer subsystem(s) are also configured for automatically determining parameter(s) of measurement(s) performed in first and second subsets of the ROIs during the metrology process with the measurement subsystem based on portions of the design for the specimen located in the first and second subsets of the ROIs, respectively. The parameter(s) of the measurement(s) performed in the first subset are determined separately and independently of the parameter(s) of the measurement(s) performed in the second subset.

Automated pattern fidelity measurement plan generation

Methods and systems for determining overlay error between different patterned features of a design printed on a wafer in a multi-patterning step process are provided. For multi-patterning step designs, the design for a first patterning step is used as a reference and designs for each of the remaining patterning steps are synthetically shifted until the synthetically shifted designs have the best global alignment with the entire image based on global image-to-design alignment. The final synthetic shift of each design for each patterning step relative to the design for the first patterning step provides a measurement of relative overlay error between any two features printed on the wafer using multi-patterning technology.

Determining multi-patterning step overlay error

One embodiment relates to a method of inspecting an array of cells on a substrate. A reference image is generated using a cell image that was previously determined to be defect free. A reference contour image which includes contours of the reference image is also generated. The reference contour image is used to detect defects in the array of cells on the substrate. Another embodiment relates to a system for detecting defects in an array on a substrate. Other embodiments, aspects and features are also disclosed.

Growing community of inventors

Milpitas, CA, United States of America

Thanh Huy Ha