Use this URL to cite or link to this record in EThOS:
Title: Automatic registration of 3D models to laparoscopic video images for guidance during liver surgery
Author: Robu, Maria Ruxandra
ISNI:       0000 0004 9352 9244
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Laparoscopic liver interventions offer significant advantages over open surgery, such as less pain and trauma, and shorter recovery time for the patient. However, they also bring challenges for the surgeons such as the lack of tactile feedback, limited field of view and occluded anatomy. Augmented reality (AR) can potentially help during laparoscopic liver interventions by displaying sub-surface structures (such as tumours or vasculature). The initial registration between the 3D model extracted from the CT scan and the laparoscopic video feed is essential for an AR system which should be efficient, robust, intuitive to use and with minimal disruption to the surgical procedure. Several challenges of registration methods in laparoscopic interventions include the deformation of the liver due to gas insufflation in the abdomen, partial visibility of the organ and lack of prominent geometrical or texture-wise landmarks. These challenges are discussed in detail and an overview of the state of the art is provided. This research project aims to provide the tools to move towards a completely automatic registration. Firstly, the importance of pre-operative planning is discussed along with the characteristics of the liver that can be used in order to constrain a registration method. Secondly, maximising the amount of information obtained before the surgery, a semi-automatic surface based method is proposed to recover the initial rigid registration irrespective of the position of the shapes. Finally, a fully automatic 3D-2D rigid global registration is proposed which estimates a global alignment of the pre-operative 3D model using a single intra-operative image. Moving towards incorporating the different liver contours can help constrain the registration, especially for partial surfaces. Having a robust, efficient AR system which requires no manual interaction from the surgeon will aid in the translation of such approaches to the clinics.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available