Intelligent Cancer Detection System
Utpol Kanti Das, Juel Sikder, Umme Salma, Afsah Anwar
Abstract
Histopathology and MRI produce different types of images and a single procedure to detect cancerous cells from both of them is considered tedious and time-consuming. Therefore, this study proposes a strategy that can perform segmentation, classification and detection over numerous categories of cancer cells. This experimentation demands the preprocessing to intensify the cells’ pictorial characteristic and to segment the intended cancerous cell k-means algorithm applied. The Convolutional Neural Network (CNN) classifier being considered as the best method classifying the test image using the feature extraction method on the filtered cell. Considering the system detects cancer, different morphological operations are applied on the extracted cancer region, and finally, we calculated the percentage of the cancerous area. The MATLAB environment has been used on Histopathology and MRI images of the infected cell from various authentic datasets of various cancers mainly Brain, Breast, Leukemia and Lung cancer. The proposed arrangements of methods give improved accuracy as compared to the other existing methods.