Mamogram Image Classification Using Extreme Learning Machine

Objective: To make early diagnosis of breast cancer from mammogram image. Methods/Statistical analysis: To avoid the misdiagnosis, we proposed a system to sort the suspicious masses from the mammogram image by using Extreme learning machine algorithm. The ELM based classifier is used to classify the input data as malignant and benign classes with the abnormal class. The effectiveness of the ELM algorithm is superior to the other existing algorithms for mammogram classification problems with its reduced training time and classification accuracy. Findings: We provide an optimistic method for binary class classification of mammograms using extreme learning machine algorithm. Mammography is a technique which is preferred for early diagnosis of breast cancer. On the other hand, in most cases, it is not easy to differentiate benign and malignant tumor without biopsy, hence misdiagnosis is always possible. The machine learning algorithm provides high accuracy than other techniques and also the execution time is very low when compared to normal diagnosis. The existing methods are very slow compared to this proposed technique. The input images are the mammogram image and the segmentation, preprocessing is performed to remove the noises present. Application/Improvements: The main application of the system is the early diagnosis of cancerous cell present and also classifies the normal and abnormal images.