Method and system for document image classification
AssigneeXerox Corporation
InventorsAsghar Nafarieh, Alejandro E. Brito
Life of this patent
Estimating the term…
Abstract
A method of classifying an input image includes the initial steps of labeling an input image in accordance with a class and extracting at least one connected component from the input image. The method also includes the steps of calculating at least one feature of the input image and generating a model based on the at least one calculated feature. The method also includes the steps of repeating at least one of the previous steps for at least one other input image and comparing the at least one other input image with the model. The at least one other input image is classified in accordance with the class of the model if the at least one calculated feature of the at least one other input image is substantially similar to that of the model.
Claims20
A method of classifying an input image, comprising:
- labeling an input image in accordance with a class;
- extracting at least one connected component from the input image;
- calculating at least one feature of the input image based on the at least one extracted connected component;
- generating a model based on the at least one calculated feature, the model corresponding to the class of the input image;
- repeating at least one of the previous steps for at least one other input image;
- and
- comparing the at least one other input image with the model, wherein the at least one other input image is classified in accordance with the class of the model if the at least one calculated feature of the at least one other input image is substantially similar to that of the model.
A method of classifying an input image, comprising:
- labeling an input image in accordance with a class;
- extracting at least one connected component from the input image;
- calculating at least one feature of the input image based on the at least one extracted connected component;
- repeating at least one of the previous steps for at least one other input image;
- clustering input images of the same class into one of at least one sub-class;
- aggregating the at least one calculated feature of the at least one clustered input image of the sub-class;
- and
- generating a model of the sub-class based on the at least one aggregated feature of the at least one clustered input image, wherein an input image having at least one calculated feature substantially similar to the at least one aggregated feature of the model is classified in accordance with the class of the at least one clustered input image of the sub-class.
A method of classifying an input image, comprising:
- bounding at least one target of at least one input image;
- labeling the at least one bounded target in accordance with a class of the at least one input image;
- extracting at least one connected component from the at least one bounded target;
- calculating at least one feature of the at least one bounded target based on the at least one extracted connected component;
- generating a model of the at least one bounded target based on the at least one calculated feature, wherein the model is labeled in accordance with the class of the at least one input image to which the label of the at least one bounded target corresponds;
- extracting at least one connected component from at least one other input image;
- calculating at least one feature of the at least one other input image based on the at least one extracted connected component;
- and
- comparing the at least one other input image with the model, wherein the at least one other input image is labeled in accordance with the class of the model if the at least one calculated feature of the at least one other input image is substantially similar to that of the model.
Dependencies are inferred from the claim text.