Performance evaluation of optimized convolutional neural network mechanism in the detection and classification of ovarian cancer.

Female mortality is frequently caused by ovarian cancer (OC). Because of its late detection, ovarian cancer seems to have a low survival rate, and new techniques are required for its early identification. One of the more prevalent gynecologic cancers is ovarian cancer. The various diagnoses of ovarian cancer depend on the efficient classification of the various forms. Patients with ovarian tumours require accurate diagnosis. When compared to a deep convolutional neural network, previous neural networks are an outmoded technology that offers fewer characteristics, which demonstrates that deep convolutional layers supply essential and healthy features. To get over these limitations, ovarian tumours are identified using the krill herd optimization-based convolutional neural network (KHO-CNN) mechanism, a novel optimized deep neural network approach. The system analyses datasets related to ultrasound-detected ovarian cancer. The obtained real-world ultrasound images of ovarian cancer also contain additional noise, which is removed using a Wavelet Transform. An enhanced KHO model has been used in the segmentation process. Features were extracted by use of a local binary pattern. Ovarian tumours are classified as benign, malignant, or normal by the KHO-CNN. To identify ovarian cancers using deep learning techniques that utilize optimised convolutional neural networks, this model was developed and utilised with a set. [ABSTRACT FROM AUTHOR]