Information processing apparatus that performs machine learning of learning model, learning method, and storage medium
AssigneeCanon Kabushiki Kaisha
InventorHayato Oura
Life of this patent
Estimating the term…
Abstract
An apparatus that performs machine learning of a learning model that can increase a rate of outputting a block dividing pattern as a correct answer. A CPU of the apparatus determines a dividing pattern output from a learning model that outputs, using an image for learning as an input thereto, a dividing pattern obtained by dividing the image, and a weight associated with a similarity with teacher data when performing machine learning of the learning model. The CPU performs machine learning of the learning model using a loss function weighted by the determined weight.
Claims18
An apparatus comprising:
- at least one processor;
- and
- at least one memory coupled to the at least one processor, the at least one memory having instructions that, when executed by the at least one processor, perform the operations as:
- a determination unit configured to determine a weight based on a similarity between a dividing pattern output from a learning model that outputs, using an image for learning as an input thereto, a dividing pattern obtained by dividing the image, and teacher data being used for performing machine learning of the learning model;
- and
- a learning unit configured to:
- generate a learned model by performing machine learning of the learning model using a loss function weighted by the determined weight,
- wherein the similarity is determined by counting a number of sections of the dividing pattern output from the learning model, which differ from a dividing pattern of the teacher data;
- and encode an unknown image forming a moving image using a block dividing pattern output from the learned model, such that a probability that the block dividing pattern output from the learned model is similar to the output teacher data is increased.
A method comprising:
- determining a weight based on a similarity between a dividing pattern output from a learning model that outputs, using an image for learning as an input thereto, a dividing pattern obtained by dividing the image, and teacher data being used for performing machine learning of the learning model;
- generating a learned model by performing machine learning of the learning model using a loss function weighted by the determined weight,
- wherein the similarity is determined by counting a number of sections of the dividing pattern output from the learning model, which differ from a dividing pattern of the teacher data;
- and encoding an unknown image forming a moving image using a block dividing pattern output from the learned model, such that a probability that the block dividing pattern output from the learned model is similar to the output teacher data is increased.
A non-transitory computer-readable storage medium storing a program for causing a computer to execute a method, the method comprising:
- determining a weight based on a similarity between a dividing pattern output from a learning model that outputs, using an image for learning as an input thereto, a dividing pattern obtained by dividing the image, and teacher data being used for performing machine learning of the learning model;
- generating a learned model by performing machine learning of the learning model using a loss function weighted by the determined weight,
- wherein the similarity is determined by counting a number of sections of the dividing pattern output from the learning model, which differ from a dividing pattern of the teacher data;
- and encoding an unknown image forming a moving image using a block dividing pattern output from the learned model, such that a probability that the block dividing pattern output from the learned model is similar to the output teacher data is increased.
Dependencies are inferred from the claim text.