MICCAI 2016 Daily - Tuesday

Segmentation of Empty Catheters in PCI Procedure MICCAI Daily: What is your position, Ketan? Ketan: I am a PhD student with GE Healthcare with Dr. Régis Vaillant and University of Paris-Est with Professor Laurent Najman and Professor Jean Cousty. MICCAI Daily: I know that today you had a presentation. Can you tell us about the work that you presented? Ketan: It’s about surgical process modeling in cardiac X-ray fluoroscopic images. MICCAI Daily: What is the medical interest of your work? Ketan: It’s a nice question because it is a very long term interest. It’s an explorative work that we have been doing. The acquisition system will know what the doctor is doing and what part of the procedure the doctor is in at the present moment. The system will know what action is being taken in a real time basis. This will enable easy interaction of the user, which is the doctor, and the machine. It will also enable launching of some applications automatically which will be more modern. MICCAI Daily: What kind of image processing techniques did you use in order to create your model? Ketan: I work with a very cool method and algorithms in image processing. I use mathematical morphology. This is the framework of connected operators that I used for the segmentation task. It is a unique framework for me because it’s a single framework that will help me to segment multiple objects. I don’t have to adopt different strategies for segmenting different objects. It’s a unifying framework. Mathematical morphology provides really nice tools for that. My professors have been working on mathematical morphology for the last 25 years and this is a lab which specializes in mathematical morphology. MICCAI Daily: What are your next steps in this project? Ketan: It is part of my PhD so my next steps are to segment more objects, put the different objects together, to find the distance between them, spatially and temporarily, and use this distance to level the action that’s performed in the sequence. This is something cool because it hasn’t been tried before. No one has tried the semantic analysis of the PCI procedure. CVPR Daily: Thursday Presentation 19 MICCAI Daily: Tuesday Ketan Bacchuwar