Breast Cancer Classification Using Customized Convolution Neural Network

Cancer involves the unreviewed proliferation of each bodily cell, capable of quickly spreading to any part of the body. Breast cancer remains a concern, in health and it requires for the development of effective methods to detect it early. This study used TensorFlow 2.15 along with a CNN to improve the accurate identification of breast cancer detection. By analyzing dataset which is Breast Cancer Wisconsin dataset, proposed CNN model, which incorporates Conv1D layers, batch normalization and dropout techniques achieved an accuracy rate of 96% in classifying tumors as either malignant or benign. In the research analysis section, proposed work analyzes the epochs for this model with visualizations which is accuracy and loss graph for training and validation, it shows training accuracy is rising at end of the 50 epoch with 98.02% accuracy and validation accuracy is also rising after 10 epochs with 91.23% to 50 epoch with 97.37% accuracy, it shows the high accuracy of this proposed model. In the Loss Graph training loss is decreasing from 10 epochs with 0.0981 MSE to 50 epoch with 0.0522 MSE and validation loss is decreased at 50 epoch with 0.1497 MSE. This research the possibility to acquire knowledge in medical image analysis and its ability to enhance diagnosis accuracy for breast cancer cases. The outcome emphasize the efficiency of the proposed strategy and pave the way for advancements, in utilizing CNNs for critical healthcare applications.