The mathematical model for the differential diagnosis of endoscopic gastric ulcerative lesions

• DOI: 10.3760/cma.j.issn.1007-5232.2015.03.012
• Publication Date: 2015-03-20
• Authors: Hai‐Yan Tao; Peng Li; Ping Zhou; Shutian Zhang
• Keywords: Perimeter; Roundness (object)

Objective To establish mathematical model for the differential diagnosis of endoscopic gastric ulcerative lesions with help image processing software and statistical analysis method.

Methods The microscopic images 25 ulcers peptic cancers were collected from January 2013 to December 2013.Then Photoshop CS4 was used representative measurements gathering(area, perimeter, roundness, height, width, gray scale, involving density, color, lightness, pixels). Then principal component analysis, scatter plot cluster performed by SPSS 16.0.

Results According results Z1 showed a strong positive correlation area, cumulative pixel, which represented shape statistics ulcer. And Z2 color On Scatter plot, cancer stronger Z2. Clustering that majority patients, there pattern in its incidence. In ulcer diagnostic models, effective information should include area≤10 000; perimeter≤250; height≤90; width≤60; scale≥90; density≤190 color≥110; lightness≥110(unit: pixel). cancers, area≥38 000, perimeter≥690, height≥300, width ≥426, scale≤100, density≥14 000 color≤125, lightness≤130(unit: pixel).

Conclusion Among data, brightness, pixels are important indices, can contribute lesions. The established based on ten indices be applied predict diagnose lesions.

Key words: Principal analysis; Factor analysis; Cluster analysis; Endoscopy; Gastric ulcer; Gastric