An Enhanced Technique for Lymphoblastic Cancer Detection Using Artificial Neural Network
Savita Dumyan, Ankush Gupta · International journal of advanced research in computer science and electronics engineering · 2017
This paper proposes a technique for increasing the accuracy of lymphoblastic cancer/cell detection in blood samples. The proposed technique utilizes the microscopic images of blood samples to visualise and identify the lymphoblastic cancer cells. These microscopic images are first enhanced using several image processing steps and subsequently, several features are extracted by examining the changes in texture, shape, color, and statistical behaviour of these enhanced images; thereafter, the artificial neural network is trained based on these extracted features in order to detect these cancer cells with high accuracy. Accordingly, the trained artificial neural network classifier categories the image into two classes, namely; normal and abnormal White Blood Cells by using an appropriate set of features and hence, detecting the lymphoblastic cancer/cells. For proof of concept, 36 microscopic images of blood samples are tested using the proposed technique and an overall accuracy of 97.8% is achieved.