Ovarian Cancer Detection and Diagnosis Using Deep Learning Structure
K. Bala Maheswari, S. Gomathi
Abstract
Among cancers affecting women, ovarian cancer ranks sixth in terms of frequency of mortality. The first symptoms are often vague and changeable, which makes it difficult to distinguish between stages III and IV. Subjective interpretation, inconsistency among observers, and lengthy testing periods are some of the limitations of current diagnostic methods. These methods include imaging tests, biopsies, and biomarkers. In order to extract features, the proposed method makes use of the SIFT technique. An in-depth investigation of every object is possible by extracting its various unique qualities and interesting elements. By making use of the fitness function, the extracted feature is optimized using the genetic algorithm. Using an architecture of convolutional neural networks, accurate cancer of the ovary detection and staging was achieved. Achieving 97% accuracy, the CNN classifier outperforms the SVM's 84 % performance.