Classification of oral cavity cancer using linear discriminant analysis (LDA) and principal component analysis (PCA)

Oral cavity cancer ranks as the fourth most frequent cancer among men and eighth for women, which has a significant effect on human health. The diagnosis of oral cavity cancer is an expensive and inconvenient examination. This work aims to propose an effective method for identifying oral cancers at an earlier stage using linear discriminant analysis (LDA) with the texture content of an image at the gray level and principal component analysis (PCA). The experimental results of this work indicate that statistical image analysis can be used as a complementary tool in the diagnosis of oral lesions. The PCA features with LDA produced a classification of 98.2(±3.9) % and 95.4(±4.9) % for Leukoplakia (potentially precancerous) and Lichenoid (harmless) lesions, respectively, with the most errors found to be false positive errors.