With the support of the Dutch EFRO (European Regional Development Fund) and SNN (Northern Netherlands Alliance) / VIA 2020 (Accelerated Innovation), Real AI B.V. has set out to do a breast cancer detection project. CytoAI is our cutting edge healthcare AI project with the goal of developing a POC ( proof of concept) so that it can be tested and validated against previously unseen breast cancer images.
Breast cancer is the most common cancer in women. Demographic changes, different dietary patterns, and life habits are prompting even younger women to start scheduling routine checks for tumors. As the digitalization of society continues, we also expect that most women will be encouraged to use such digital solutions instead of going through complicated and cumbersome procedures to get a diagnosis. It is already a difficult decision for women, and we believe our solution will not only help change the way we look at breast cancer diagnostics, but also how doctors and oncologists will be able to embrace intelligent tools built and developed by AI.
CytoAI is our pioneering healthcare AI project with the goal of developing aPOC ( proof of concept) so that it can be tested and validated against previously unseen breast cancer images. In the envisioned final product, this data can be easily uploaded by both doctors and patients, after which the tool will immediately classify the image to see if there are signs of malignancy or benignity. With CytoAI, the diagnosis of breast cancer will become a simple and easy exercise, while statistically, the number of misdiagnoses or missed diagnoses will be drastically reduced.
In the following four phases, we intend to build a compelling and high-impact POC (proof of concept) that is reliable and reproducible.
collect, understand and process data
In this phase we collected relevant breast cancer data, after pre-processing and performing thorough data exploration analysis on the pre-processed data (EDA), we developed a dataset.
model choices and development
Here we conducted a thorough research-driven analysis on the available AI platforms, most appropriate AI Algorithms, most optimal model choices and finally the best-in-class AI model and algorithm development strategy and planning.
optimize and deploy ai model
Here we optimised the AI model with latest and cutting-edge techniques available in the industry. We also deployed some of our own toolkit to further enhance the model performance.
Within this VIA project, we will demonstrate the software code for the working principles of of the software prototype to be developed (proof-of-concept software).