Multiclass Stomach Diseases Classification Using Deep Learning Features Optimization

In the area of medical image processing, stomach cancer is one most important cancers which need to be diagnose at early stage. this paper, an optimized deep learning method presented for multiple disease classification. The proposed work in few steps—preprocessing using fusion filtering images along with Ant Colony Optimization (ACO), transfer learning-based features extraction, optimization extracted nature-inspired algorithms, and finally optimal vectors classification Multi-Layered Perceptron Neural Network (MLNN). feature extraction step, pre-trained Inception V3 utilized retrained on selected infection classes step. Later on, activation function applied Global Average Pool (GAP) extraction. However, are through two different algorithms—Particle Swarm (PSO) dynamic fitness Crow Search Algorithm (CSA). Hence, both methods’ output fused by a maximal value approach classified vector MLNN. Two datasets used evaluate method—CUI WahStomach Diseases Combined dataset achieved average accuracy 99.5%. comparison existing techniques, it shown that shows significant performance.