An image processing method for extracting and normalizing salient features from images with constant feature changes

Abstract

The invention provides an image processing method for extracting significant features from a feature constant image and normalizing the significant features. The method mainly comprises the followingsteps of: reading a single chromosome subgraph subjected to image enhancement and straightening, extracting the overall outline of a chromosome in each subgraph, extracting a connected region representing a deep zone in the outline of the chromosome, determining the center and the axial width of each deep zone region, and normalizing the deep zone regions into a complete strip. The main parts respectively comprise a series of image processing algorithm methods, which comprises threshold segmentation, contour filling, region marking, distance transformation and the like, so as to finally achieve the purpose of accurately extracting and normalizing the salient strips. Results show that the band arrangement modes of the chromosomes of the same type are more similar after normalization and aremore similar to those of a standard mode diagram.

Description

technical field [0001] The invention relates to an image processing method for extracting and normalizing salient features from images with constant feature changes, belonging to the field of image processing. Background technique [0002] Chromosomal karyotype analysis plays a vital role in the detection of genetic diseases. After the cell samples have gone through the steps of culturing, harvesting, film production and scanning imaging, the karyotype analyst can perform karyotype analysis on the chromosome image. [0003] Generally, only one chromosome of a cell is focused on in a chromosome image, and all chromosomes in the cell are scattered and arranged. The banding pattern of each chromosome is an important basis for karyotype analysis. With the rapid development of artificial intelligence, computer-aided chromosome karyotype analysis technology has also made significant progress, and can even identify the type of chromosome or detect the possibility of chromosomal a...

AI Technical Summary

Problems solved by technology

However, due to the fact that many aspects of film production are done manually and the chain structure of the chromosome itself, the band arrangement and coloration degree of the chromosome often have a certain degree of randomness in the imaging (including the bending of the band, sometimes no, sometimes Deep and shallow, etc.)

The randomness in the arrangement of chromosome bands and the degree of coloration in the image often causes a certain degree of error in computer recognition and detection

Images