We are happy to share that new algorithm developed within the PANCAIM projcet has been published!
The AI model developed and trained by researcher Pierpaolo Vendittelli – PhD Student at Radboud University Medical Center – is available on the Grand Challenge platform now!
This algorithm applies a three-steps pipeline for segmenting Pancreatic Ductal Atenocarcinoma in pancreatic whole-slide images – two deep learning networks and a script for morphological operations:
- First the input image is used to segment the tissue from the background (this is a required step to drastically reduce the inference time)
- The input image is then used together with the tissue mask to segment the healthy epithelium (in green) and the tumor epithelium (in yellow), as shown in the second image.
- A convex hull is then created around the tumor epithelium and a tumor bulk mask is generated (third image).
The time needed by the algorithm to produce result is approximately 9 minutes.
We are looking forward to make this algorithm available to all our collaborators as soon as possible for further testing and validation.
