Tumor Segmentation and Classification Using Machine Learning Approaches

Abstract

Medical image processing has recently developed progressively in terms of methodologies and applications to increase serviceability in health care management. Modern medical image processing employs various methods to diagnose tumors due to the burgeoning demand in the related industry. This study uses the PG-DBCWMF, the HV area method, and CTSIFT extraction to identify brain tumors that have been combined with pancreatic tumors. In terms of efficiency, precision, creativity, and other factors, these strategies offer improved performance in therapeutic settings. The three techniques, PG-DBCWMF, HV region algorithm, and CTSIFT extraction, are combined in the suggested method. The PG-DBCWMF (Patch Group Decision Couple Window Median Filter) works well in the preprocessing stage and eliminates noise. The HV region technique precisely calculates the vertical and horizontal angles of the known images. CTSIFT is a feature extraction method that recognizes the area of tumor images that is impacted. The brain tumor and pancreatic tumor databases, which produce the best PNSR, MSE, and other results, were used for the experimental evaluation.