A Novel Approach for Increased Convolutional Neural Network Performance in Gastric-Cancer Classification Using Endoscopic Images

Gastric cancer is the third-most-common cause of cancer-related deaths in world. Fortunately, it can be detected using endoscopy equipment. Computer-aided diagnosis (CADx) systems help clinicians identify from gastric diseases more accurately. In this paper, we present a CADx system that distinguishes and classifies pre-cancerous conditions, such as polyps, ulcers, gastritis, bleeding. The uses deep-learning model, Xception, which involves depth-wise separable convolutions, to classify non-cancers. proposed method consists two steps: Google's AutoAugment for augmentation simple linear iterative clustering (SLIC) superpixel fast robust fuzzy C-means (FRFCM) algorithm image segmentation during preprocessing. These approaches produce feasible distinguishing classifying cancers other diseases. Based on biopsy-supported ground truth, performance metrics area under receiver operating characteristic curve (i.e. Az) are measured test sets. classification results, Az model 0.96, 0.06 up 0.90 original data. Our methods fully automated without manual specification region-of-interests with random selection images training. This methodology may play crucial role selecting effective treatment options need surgical biopsy.