Method and apparatus with optimization and prediction for image segmentation
AssigneeSamsung Electronics Co., Ltd.
InventorsPingjun LI, Yi ZHOU, Byung In YOO, Hana LEE, Hui ZHANG
Life of this patent
Estimating the term…
Abstract
A processor-implemented method includes: determining a probability that a pixel of an input image belongs to each of a plurality of preset categories; and determining a category of the pixel to be a category corresponding to either one or both of a plurality of category areas and a category determined based on the probability that the pixel belongs to each of the preset categories, based on a result of comparing, to a preset threshold value, a probability that the pixel belongs to the category corresponding to the category areas.
Claims19
A processor-implemented method, comprising:
- determining a probability that a pixel of an input image belongs to each of a plurality of preset categories;
- and
- determining a category of the pixel to be a category corresponding to either one or both of a plurality of category areas and a category determined based on the probability that the pixel belongs to each of the preset categories, based on a result of comparing, to a preset threshold value, a probability that the pixel belongs to the category corresponding to the category areas.
A processor-implemented method, comprising:
- determining a multi-category image;
- and
- optimizing a first pixel classifier using the multi-category image,
- wherein a pixel of the multi-category image corresponds to two or more categories in a plurality of category areas.
An apparatus, comprising:
- one or more processors configured to:
- determine a probability that a pixel of an input image belongs to each of a plurality of preset categories;
- and
- determine a category of the pixel to be either one or both of a category corresponding to a plurality of category areas and a category determined based on the probability that the pixel belongs to each of the preset categories, based on a result of comparing a probability that the pixel belongs to the category corresponding to the category areas to a preset threshold value.
An apparatus, comprising:
- one or more processors configured to:
- determine a multi-category image;
- and
- optimize a first pixel classifier using the multi-category image,
- wherein a pixel of the multi-category image has at least two categories in a plurality of category areas.
A processor-implemented method, comprising:
- obtaining a single category image in which each pixel is determined to correspond to a single category among preset categories;
- determining, based on the single category image, a multi-category image in which at least one pixel is determined to correspond to a plurality of categories among the preset categories;
- and
- optimizing a first pixel classifier based on the multi-category image.
Dependencies are inferred from the claim text.