Eshika Saxena

HemaCam: A Mobile Phone Microscopy System for Automated Screening of Hematological Diseases Using Computer Vision Techniques and Convolutional Neural Networks
Eshika Saxena

Eshika Saxena

Bellevue, WA

College Freshman, Inducted 2019

Harvard University, Computer Science

 

Hematological diseases affect millions of people worldwide. Early diagnosis is key to treating/mitigating these diseases, but diagnosis is costly, time-intensive, and requires trained medical professionals. Automated or in-home screening can greatly improve the prognosis and treatment of these diseases. My research focused on developing a low-cost, smartphone-based, solution that instantly screens for hematological diseases. I successfully designed a clip-on smartphone microscope attachment that captures blood cell images from a blood-smear. I applied computer vision techniques to segment the individual cells and trained deep-learning models to automatically classify blood cell morphology, detect abnormalities, and screen for diseases.