Performance Analysis of Local Binary Pattern Features with Eagle Optimization Algorithm based CNN Classifier for Breast Cancer Detection from Ultrasound Images

Breast cancer detection at early stages is very much needed as it is one of the most prevalent cancer related death causing cancers among females. This paper poses an automated breast cancer detection based on analyzing ultrasound images using modern deep learning and optimization algorithms. We implement Local Binary Pattern feature extraction in addition to the Cuckoo Search Algorithm and Eagle Optimization Algorithm based CNN to improve the accuracy of the classification. The Eagle Optimization Algorithm on CNN classifier $\mathbf{97.89\%}$ results higher than the Cuckoo Search Algorithm on CNN classifier 95.46%. These results highlight the importance of good feature extraction and selection in enhancing the diagnosis.