Method of producing machine learning model, and copying apparatus
AssigneeSeiko Epson Corporation
InventorsYo Kawano, Toshifumi Sakai
Life of this patent
Estimating the term…
Abstract
A method of producing a machine learning model includes collecting, as training data, data including an image of a document and a size of the document or a size of an actually copied output sheet from a copying apparatus that copies the document on the output sheet, and producing a model for determining whether or not a mark included in the image of the document is identifiable by a user on the document or the copied output sheet based on the training data.
Claims5
A method of producing a machine learning model, the method comprising:
- collecting, as training data, data including an image of a document and a size of the document;
- and
- producing a model that indicates a correspondence between an image and a size of an output sheet required for printing the image such that a mark included in the image is identifiable by a user or a relationship between an input that includes the image and a predetermined size and an output that includes a value indicating that the mark included in the image is identifiable or not by the user on an output sheet of the predetermined size through machine learning based on the training data.
A copying apparatus that reads a document, and prints an image of the document on an output sheet, the apparatus comprising:
- a processor that outputs, as training data, data including the image of the document and a size of the document or a size of the output sheet, and performs machine learning using the training data,
- the processor obtaining a model produced by the machine learning, the model indicating a correspondence between an image and a size of an output sheet required for printing the image such that a mark included in the image is identifiable by a user or a relationship between an input that includes the image and a predetermined size and an output that includes a value indicating that the mark included in the image is identifiable or not by the user on an output sheet of the predetermined size, and
- the processor determining whether or not a mark included in the image of the document is identifiable by a user on the output sheet by using the model before printing the image of the document on the output sheet, and controlling the printing based on the determination result when it is assumed that the image of the document is printed on an output sheet of a predetermined size.
A method of producing a machine learning model, the method comprising:
- collecting, as training data, data including an image of a document and a size of an output sheet from a copying apparatus that reads the document and prints the image of the document on the output sheet;
- and
- producing a model that indicates a correspondence between an image and a size of an output sheet required for printing the image such that a mark included in the image is identifiable by a user or a relationship between an input that includes the image and a predetermined size and an output that includes a value indicating that the mark included in the image is identifiable or not by the user on an output sheet of the predetermined size through machine learning based on the training data.
Dependencies are inferred from the claim text.