Classification with Random Forest Based on Local Tangent Space Alignment and Neighborhood Preserving Embedding for MSER features: MSER_DFT_LTSA-NPE_RF

— In this article, 180 gastric images taken with Light Microscope help are used. Maximally Stable Extremal Regions (MSER) features of the images for classification has been calculated. These MSER features have been applied Discrete Fourier Transform (DFT) method. High-dimensional of these MSER-DFT feature vectors is reduced to lower-dimensional with Local Tangent Space Alignment (LTSA) and Neighborhood Preserving Embedding (NPE). When size reduction process was done, properties in 5, 10, 15, 20, 25, 30, 35, 40, 45, and 50 dimensions have been obtained. These low-dimensional data are classified by Random Forest (RF) classification. Thus, MSER_DFT_LTSA-NPE_RF method for gastric histopathological images have been developed. Classification results obtained with these methods have been compared. According to the other methods, classification results for gastric histopathological images have been found to be higher.