Symbol Recognition Method
AssigneeCompal Electronics, Inc.
InventorsMing-Hung Lin, Chih-Wei Lin
Life of this patent
Estimating the term…
Abstract
A symbol recognition method for increasing the symbol recognition speed by the processor includes the steps of obtaining a pixel density value and an aspect ratio of the symbol image. Then, the center-point value and the corresponding radius value are obtained by a partitional clustering algorithm. In sequence, a pixel density value and the aspect ratio of the under recognized image are obtained, and compared with the values of the symbol image to determine whether the under recognized image is a single symbol image or not.
Claims15
A symbol recognition method for increasing the symbol recognition speed by the processor, comprising:
- acquiring a plurality of symbol images;
- obtaining a pixel density value and an aspect ratio of the symbol image;
- obtaining a plurality of center-point values and a plurality of radius values of the symbol images through processing the pixel density values and the aspect ratios of the symbol images based on a partitional clustering algorithm;
- acquiring an under recognized image;
- obtaining a pixel density value and an aspect ratio of the under recognized image;
- and
- comparing a distance between a recognized point value determined by the pixel density value and the aspect ratio of the under recognized image and one of the center-point value with the radius value corresponding to the center-point value;
- wherein determining the under recognized image is a single symbol image when the distance is smaller than the radius value corresponding to the center-point value, and
- determining the under recognized image is a connected symbol image when the distance is larger than the radius value corresponding to the center-point value.
A symbol recognition method for increasing the symbol recognition speed by the processor, comprising:
- acquiring a plurality of symbol images;
- obtaining a pixel density value and an aspect ratio of the symbol image;
- obtaining a plurality of center-point values and a plurality of radius values of the symbol images through processing the pixel density values and the aspect ratios of the symbol image based on a partitional clustering algorithm;
- acquiring a under recognized image;
- obtaining a pixel density value and an aspect ratio of the under recognized image;
- comparing a distance between a recognized point value determined by the pixel density value and the aspect ratio of the under recognized image and one of the center-point value with the radius value corresponding to the center-point value;
- determining the under recognized image as a single symbol image when the distance is smaller than the radius value corresponding to the center-point value;
- analyzing the under recognized image by a symbol edge detection when the distance is larger than the radius value corresponding to the center-point value;
- and
- dividing the under recognized image when the under recognized image being determined as a connected symbol image according to the results of the symbol edge detection.
Dependencies are inferred from the claim text.