Breast Cancer Cell Detection using FCM and Prediction using UNET based Deep Convolutional Neural Network
L. K. Sravanthi Potti, S. Maruthuperumal · 2024
Breast cancer is the primary cause of death in women, resulting in a relatively high survival rate. We employ FCM to segment and detect cancerous cells in breast tissue images. UNET architecture is used for feature extraction from input image. We then use a deep CNN model to predict and classify the malignancy, which enhances diagnostic accuracy and facilitates early detection and treatment planning. This is achieved by predicting breast cancer at an earlier stage. Breast cancer prediction and diagnosis are extremely important for women's healthy lives. To predict breast cancer, we must use a recent and high-accuracy algorithm, ensuring satisfactory medical analysis and performance. In the analysis of breast cancer, we will consider two key points: the first is segmentation, and the second is classification. Both tasks are critical to better understanding breast cancer. For early-stage identification of breast cancer, ML algorithms are more commonly used, but machine learning algorithms have a few limitations, like low accuracy for complex pathology datasets and a longer training time. This paper employs an UNET based deep learning algorithm known as UNET- deep CNN (Convolutional Neural Network) to circumvent these limitations. We found that the UNET based deep CNN algorithm outperforms the state-of-the-art breast cancer prediction technique SVM (Support Vector Machine). We calculate the performance of Fuzzy C Means with UNET based deep CNN using performance metrics. The accuracy obtained by Fuzzy C Means with the UNET based deep CNN algorithm for 20% test data is 99.23%. For a given dataset with we observe that the proposed model outperforms the state-of-the-art SVM for a given dataset with breast cancer images.