Automated detection of toxicity in the blood samples of dye factory workers

Textile processing units largely employ azo dyes which are derivatives of benzene. Toxic sewage discharged from textile industries contains azo dyes. They affect soil fertility, water resources, marine organisms and the ecosystem. Benzedine is responsible for skin irritation and also increases the toxicity in blood and bone marrow which affects the production of blood cells and hence leads to anemia. Exposure to these benzedine dyes can induce hemolytic anemia which leads to deformation of Red Blood Cells (RBCs). In this work, an automated detection of toxicity is proposed to alert the workers prior to the critical stages of cancer. The proposed system consists of a portable unit where the blood samples are analyzed and the toxic status of the dye factory workers can be transmitted from the local health centre to the nearby hospital. Microscopic images of the smeared blood samples are acquired using the digital microscope. These RBC images of normal and abnormal are segmented using Laplacian of Gaussian (LoG) segmentation. Geometric shape based features are extracted from the segmented images. Significant geometric features are chosen by conducting student's `t' test. It is observed from the results that these significant geometric features show discrimination between the normal and abnormal RBCs. This early detection of toxicity can help workers in the site to take immediate medication which could reduce the probability of adversity in their health condition.