You are here: indexmembersduygu.sarikaya

Duygu Sarikaya



Research topics

Video understanding and modeling of surgical activities Project CONDOR

Research Statement

Research areas

  • Computer vision
  • Machine learning
  • Medical imaging


[10th July 2018] My presentation for 20ème Journéedes Doctorants et Post-Doctorants du LTSI at Villejean campus (light pdf version) LTSI Day Presentation

Invited Seminar Talk at Hacettepe University (light pdf version) Hacettepe Talk

Public Datasets

ATLAS Dione Dataset for Robot-Assisted Surgery Video Understanding

ATLAS Dione dataset provides video data (86 full subject study videos (~910 action clips)) of ten surgeons from Roswell Park Cancer Institute (RPCI) (Buffalo, NY) performing six different surgical tasks on the daVinci Surgical System® (dVSS) with annotations of robotic tools per frame (for a subset of 99 action clips), actions taking place and their timestamps. It also provides information on the surgeon expertise levels based on the Dreyfus model.

For further information about the video data, annotations, tasks & subtasks, and subject demographics, please refer to the paper: Duygu Sarikaya, Jason Corso, Khurshid Guru, "Detection and Localization of Robotic Tools in Robot-Assisted Surgery Videos Using Deep Neural Networks for Region Proposal and Detection", IEEE Transactions on Medical Imaging, February 2017.
poster: Poster
pre-print PrePrint


Research Positions


Dissertation: “Exploring fusion models in computer vision for medical image computing” Advisor: Jason J.Corso

  • Master of Science (MS) - Computer Science and Engineering, University at Buffalo, Buffalo, NY, USA

Thesis: “Detection and segmentation of free blood in FAST Exam ultrasound images” Advisor: Jason J.Corso



Email :
Email :
Address : MediCIS - LTSI UMR1099 INSERM
Faculté de Médecine, Université de Rennes 1
2 avenue du Prof. Léon Bernard
CS 34317
35043 Rennes cedex - France

inserm rennes1 ltsi