With the support of the United States and India Science and Technology Endowment Foundation
(USISTEF) and Qualcomm Design in India (QDIC), engineers from Alexapath developed a portable
microscope system capable of taking glass Pap Smear slides as an input and able to provide Computer
Assisted Diagnosis (CAD). The CAD solution called AutoPap is able to analyze the digitized mobile
Whole Slide Image (mWSI)1 and determine whether or not there are abnormal cells present. The
microscope system - AiDA (Artificial intelligence Diagnostic Assistant) easily fits on any counter top
and requires no external wifi or expensive GPUs. The artificial intelligence is able to run on any device
with a processor equal to or better than the SnapDragon 820 chip.
In order to succeed, the following four tasks were completed: 1- build a robotic microscope
capable of imaging glass slides, 2- develop software for acquisition and viewing of digitized glass
slides, 3- create a Convolutional Neural Network (CNN) capable of classification and 4- deploy the
software and Ai onto a smart phone.
The end to end solution including microscope, camera and smart phone is both portable and
affordable. The potential use for the system is akin to triage. With such a large number of women in
India needing Cervical Cancer Screenings, a solution that can identify the presence of abnormal cells in
Pap Smears and has a turn around time of less than six minutes, can prove a powerful ally - allowing
human cytologists to only focus on the small number of cases needing staging. Additionally because
AutoPap functions as an “Assisted” Diagnosis which does not require De Novo classification, AutoPap
maybe able to quickly enter the market.
The worked contained in this paper is the work of Alexapath LLC with funding from the
USISTEF and QDIC. There are no conflicts of interest to report.