Image shape descriptor generation method
The invention discloses an image shape descriptor generation method, which comprises the following steps of: firstly, acquiring an image, carrying out binarization processing on the image, and distinguishing a target area from a background area; secondly, performing contour extraction on the target...
Saved in:
Main Authors | , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
27.04.2021
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention discloses an image shape descriptor generation method, which comprises the following steps of: firstly, acquiring an image, carrying out binarization processing on the image, and distinguishing a target area from a background area; secondly, performing contour extraction on the target area; carrying out distance transformation on the target area, and calculating the center point of the target area; and calculating a skeleton length graph according to the distance path of each pixel point of the target, forming a radial distance histogram from the skeleton length graph, and taking the radial distance histogram as an image shape descriptor. Compared with the prior art, the descriptor with translation, rotation, scale and illumination invariance is designed, and on the basis of the descriptor, reliable features of a target can be formed under the condition of a single sample and used for recognition and retrieval application of an image target; and for a target with hinge deformation, the shape des |
---|---|
Bibliography: | Application Number: CN202011632205 |