AI-assistive surgical instrument contaminant detection device

Unmet Need

There are approximately 700,000 healthcare-associated infections that occur in acute care hospitals each year. A proportion of these infections can be attributed to the insufficient removal of blood, tissue, and other biological contaminants from surgical instruments. One potential reason for these poorly cleaned instruments is linked to the advancement of surgical technology and the increased level of expertise required to properly clean the surgical instruments. There is a present need in healthcare to improve quality control of instrument sterilization and to alleviate the burden on sterile processing personnel.


Duke researchers have developed a medical device that uses artificial intelligence to reliably identify the presence of contaminants on surgical instruments. The device employs a convolutional neural network trained to detect rust, blood, bone, and other biological tissue on surgical tools to ensure they were properly cleaned. The device will be used in the decontamination room in the Sterile Processing Department of hospitals. After tools from operating rooms have been cleaned by technicians, the tools will be placed on the conveyor belt in the device where they will be dried and imaged from multiple angles. The screen on the device will display a green check if no contaminant is detected and a red cross along with an image of the tool with a box around the contaminated area if it is not. The device will serve as a completely automated quality control and will decrease the number of “dirty” tools passing through the sterilization process and ending up in the operating room.


  • Automated reliable identification of contaminants on surgical instruments
  • Alleviates the burden on the sterile processing personnel and eliminates human errors in surgical tool inspection
  • Will help to decrease hospital-acquired infections
Medical scissors laid out for surgery on a tray in a hospital

Duke File (IDF) Number



  • Belisle, Nicole
  • Garcia, Moriah
  • Gupta, Neil "Neil"
  • Huang, Yiying
  • Lin, Jiacheng

For more information please contact


Pratt School of Engineering