Does Proprietary Software Still Offer Protection of Intellectual Property in the Age of Machine Learning? -- A Case Study using Dual Energy CT Data
This reveals a vulnerability in intellectual property protection for medical device manufacturers, showing that compiled software may not effectively safeguard algorithms against machine learning-based reverse-engineering.
The study investigated whether proprietary software for medical image processing, specifically dual energy CT algorithms, can be reverse-engineered using machine learning, and found that mono-energetic images and iodine maps could be approximated with high accuracy (structural similarity >0.98) using only a single slice as training data.
In the domain of medical image processing, medical device manufacturers protect their intellectual property in many cases by shipping only compiled software, i.e. binary code which can be executed but is difficult to be understood by a potential attacker. In this paper, we investigate how well this procedure is able to protect image processing algorithms. In particular, we investigate whether the computation of mono-energetic images and iodine maps from dual energy CT data can be reverse-engineered by machine learning methods. Our results indicate that both can be approximated using only one single slice image as training data at a very high accuracy with structural similarity greater than 0.98 in all investigated cases.