Clinical research increasingly depends on access to high quality imaging data. Hospitals generate large volumes of images through routine clinical activity which can provide valuable insight for research programmes. 

However, many images contain identifiable text such as patient names, identification numbers or timestamps embedded directly within the image. 

Before these images can be used in research environments the sensitive information must be removed. 

Historically this process relied on manual review and redaction. Specialist staff needed to examine each image and remove identifiable text before the dataset could be approved for research use. 

This process created several difficulties. 

Manual redaction required considerable time and effort which limited the volume of images that could be prepared for research projects. The process also introduced delays in research timelines while datasets were being reviewed. 

At the same time strict information governance standards required the organisation to maintain a high level of confidence that all identifiable data had been removed. 

The organisation therefore sought a method that would improve efficiency while maintaining strong compliance with privacy requirements.