Image processing apparatus and method for style transformation
AssigneeSamsung Electronics Co., Ltd.
InventorsLei Zhang, Yehoon Kim, Chanwon Seo
Life of this patent
Estimating the term…
Abstract
Provided is an image processing method according to an embodiment, which includes: obtaining a label of a first image by inputting the first image to a recognition model; obtaining reference style data for a target reference image to which a visual sentiment label is assigned, the visual sentiment label being the same as the obtained label from among visual sentiment labels pre-assigned to reference images; generating second style data based on first style data for the first image and the obtained reference style data; and generating a second image based on the generated second style data.
Claims13
An image processing method comprising:
- obtaining a plurality of labels of a first image and probability values respectively corresponding to the plurality of labels by inputting the first image to a recognition model;
- for each of at least one candidate reference image, to which a same one or more visual sentiment label as one or more of the plurality of labels of the first image are assigned, calculating a degree of relevance to the first image by calculating an average of probability values respectively corresponding to the one or more of the plurality of labels as the degree of relevance;
- obtaining, based on the calculated degree of relevance, reference style data for a target reference image among the at least one candidate reference image;
- generating second style data based on first style data for the first image and the obtained reference style data;
- and
- generating a second image based on the generated second style data.
An image processing apparatus comprising:
- a processor;
- and
- a memory storing a recognition model and at least one program,
- wherein the processor is configured to execute the at least one program to:
- obtain a plurality of labels of a first image and probability values respectively corresponding to the plurality of labels by inputting the first image to the recognition model,
- for each of at least one candidate reference image, to which a same one or more visual sentiment label as one or more of the plurality of labels of the first image are assigned, calculate a degree of relevance to the first image by calculating an average of probability values respectively corresponding to the one or more of the plurality of labels as the degree of relevance,
- obtain, based on the calculated degree of relevance, reference style data for a target reference image among the at least one candidate reference image,
- generate second style data based on first style data for the first image and the obtained reference style data, and
- generate a second image based on the generated second style data.
Dependencies are inferred from the claim text.