US 2022/0067082 A1Published application

Inspection systems and methods including image retrieval module

AssigneeGeneral Electric Company

InventorsXiao Bian, Bernard Patrick Bewlay, Colin James Parris, Feng Xue, Shaopeng Liu, Arpit Jain, Shourya Sarcar

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method of inspecting a component using an image retrieval module includes storing an inspection image file in a memory and identifying a region of interest in the inspection image file. The method further includes accessing a database storing image files and determining feature vectors associated with the image files. The method also includes determining a hash code for each image file based on the feature vectors and classifying a subset of image files as relevant based on the hash codes. The method further includes sorting the subset of image files based on the feature vectors and generating search results based on the sorted subset of image files. The image retrieval module includes a convolutional neural network configured to learn from the determination of the feature vectors and increase the accuracy of the image retrieval module in classifying the image files.

Claims20

Dependent claims

A method of inspecting a component using an image inspection controller that includes a processor communicatively coupled to a memory and configured to operate in accordance with an image retrieval module, said method comprising:

storing, using the processor, at least one inspection image file in the memory;
identifying, using the processor, a region of interest in the at least one inspection image file;
determining, using the processor, at least one foreground feature vector associated with the region of interest in the at least one inspection image file;
accessing, using the processor, a database storing a plurality of image files;
determining, using the processor, a plurality of feature vectors associated with the plurality of image files, wherein each image file of the plurality of image files is associated with at least one feature vector of the plurality of feature vectors, and wherein the image retrieval module includes at least one convolutional neural network configured to learn from the determination of the plurality of feature vectors and increase the accuracy of the image retrieval module in classifying the plurality of image files;
determining, using the processor, at least one hash code for each image file of the plurality of image files based on the plurality of feature vectors;
classifying, using the processor, a subset of the plurality of image files as relevant based on hash codes of the plurality of image files;
sorting, using the processor, the subset of the plurality of image files based on the plurality of feature vectors;
and
generating, using the processor, search results based on the sorted subset of the plurality of image files.

7 more elements

A system for inspecting a component, said system comprising:

at least one memory comprising a database storing a plurality of image files;
and
at least one processor configured to access said at least one memory and operate in accordance with an image retrieval module, wherein said at least one processor is programmed to:
store at least one inspection image file in the memory;
identify a region of interest in the at least one inspection image file;
determine at least one foreground feature vector associated with the region of interest in the at least one inspection image file;
determine a plurality of feature vectors associated with the plurality of image files, wherein each image file of the plurality of image files is associated with at least one feature vector of the plurality of feature vectors, and wherein the image retrieval module includes at least one convolutional neural network configured to learn from the determination of the plurality of feature vectors and increase the accuracy of the image retrieval module in classifying the plurality of image files;
determine at least one hash code for each image file of the plurality of image files based on the plurality of feature vectors;
classify a subset of the plurality of image files as relevant based on the hash codes of the plurality of image files;
sort the subset of the plurality of image files based on the plurality of feature vectors for the plurality of image files;
and
generate search results based on the sorted subset of the plurality of image files.

9 more elements

An image retrieval module for identifying at least one feature in at least one image of a component, said image retrieval module configured to cause a processor to:

store at least one inspection image file in a memory;
identify a region of interest in the at least one inspection image file;
determine at least one foreground feature vector associated with the region of interest in the at least one inspection image file;
access a database storing a plurality of image files;
determine a plurality of feature vectors associated with the plurality of image files, wherein each image file of the plurality of image files is associated with at least one feature vector of the plurality of feature vectors, and wherein said image retrieval module includes at least one convolutional neural network configured to learn from the determination of the plurality of feature vectors and increase the accuracy of said image retrieval module in classifying the plurality of image files;
determine at least one hash code for each image file of the plurality of image files based on the plurality of feature vectors;
classify a subset of the plurality of image files as relevant based on the hash codes of the plurality of image files;
sort the subset of the plurality of image files based on the plurality of feature vectors for the plurality of image files;
and
generate search results based on the sorted subset of the plurality of image files.

7 more elements

Dependencies are inferred from the claim text.