ABSTRACT
Cervical cancer (CC) is the cancers in the world. In early stage, the symptoms of cervical cancer is not easily found. The initial and accurate diagnosis of cervical cancer is essential for patient outcomes. In this paper, Convolutional Neural Network (CNN) based DenseNet 121 classifier is proposed efficiently classify the images of cervical cancer. Gaussian filter is employed for preprocessing, it enhances the image quality and also reduces noise sensitivity. The adaptive thresholding is used for segment the features from the extracted cervix images. The features are selected and the raw medical images are transformed into unique features by using Gabor Wavelet Transform (GWT). The DenseNet 121classifier is proposed to improve the model prediction accuracy. For tuning the parameters of the classifier Butterfly Optimized Algorithm (BOA) is integrated. These methodologies improves the diagnostic accuracy and also contribute to better patient outcomes and more effective screening strategies, ultimately advancing cervical cancer management. The proposed approach is implemented in the python software and achieves the efficiency of 93.56%.
