North­eastern Uni­ver­sity Asso­ciate Pro­fessor of Elec­trical and Com­puter Engi­neering Jen­nifer Dy is on a quest to reveal the mys­teries of the body by pro­gram­ming com­puters to “learn” from the data they col­lect, ulti­mately helping physi­cians better treat tumors and early-​​stage skin cancers.

Dy, who has already received more than $1 mil­lion in research grants, including a pres­ti­gious National Sci­ence Foun­da­tion CAREER Award in 2004, is applying her exper­tise in machine learning, data mining, sta­tis­tical pat­tern recog­ni­tion and med­ical image analysis, to two highly inter­dis­ci­pli­nary research projects.

Dy is col­lab­o­rating with Asso­ciate Pro­fessor of Radi­ology Bin Jiang of the Uni­ver­sity of California-​​San Diego and Gre­gory Sharp of Mass­a­chu­setts Gen­eral Hospital’s radi­a­tion oncology depart­ment in an effort to build com­puter algo­rithms that can pre­dict the move­ment of a tumor in the body, such as a lung tumor that moves as a patient breathes.

If that move­ment can be tracked and recorded in a com­puter so a pre­dic­tive pat­tern emerges, physi­cians would be better able to pin­point the loca­tion of the tumor at any time, and there­fore, pin­point treatment.

The idea is that the radi­a­tion [treat­ment] can be tar­geted specif­i­cally at the cancer cells” and avoid healthy cells, she said.

In another major col­lab­o­ra­tion, Dy is working with an elec­trical and com­puter engi­neering col­league, Pro­fessor Dana Brooks, and Dr. Allan Halpern, with the der­ma­tology depart­ment of Memo­rial Sloan-​​Kettering Cancer Center, on a project to improve the imaging—and early detection—of skin cancers.

The goal of the project, “3-​​D Seg­men­ta­tion and Clas­si­fi­ca­tion of Skin Images from Con­focal Laser Scan­ning Micro­scope,” is to detect the epidermis/​dermis boundary—a hard-​​to-​​see area between skin layers where cancer develops, she said. “It’s not easy because dif­ferent skin types look dif­ferent in the images,” and skin images are often blurry.

Dy explained that she and her col­leagues are devel­oping data mining algo­rithms that will allow a com­puter to “learn” where skin cancer may lurk. They hope the com­puter algo­rithms lead to enhance­ment of image-​​tracking tech­niques, which, among other ben­e­fits, will improve the con­trast and clarity of images.

Her work on this project is sup­ported by the National Insti­tutes of Health’s Center for Inte­gra­tive Bio­med­ical Computing.

Dy grew up in the Philip­pines, and earned her bachelor’s degree at the Uni­ver­sity of the Philip­pines in elec­trical engi­neering. She went on to earn her master’s and doc­toral degrees in elec­trical and com­puter engi­neering from Purdue University.