Janatka, Miroslav Philip;
(2022)
Motion Magnification for Surgical Video.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Janatka__thesis.pdf - Other Download (10MB) | Preview |
Abstract
Accidental bleeding during a surgical procedure complicates and interrupts the workflow of the surgeon and can even lead to the operation being aborted. It can be caused by abnormal vascular anatomy that the surgeon is not expecting and unaware of as they are unable to localise from the surgical video feed alone. Image guidance can assist in localisation intraoperatively, however, requires additional equipment that adds cost to the operation. Video motion magnification has been proposed to aid vessel visualation in surgical video without the need for additional equipment beyond the visible light cameras used in endoscopes and surgical microscopes. This technique exaggerates and makes perceivable subtle variations in the video feed that are not normally perceivable to the surgeon. However, other motions in the video are also detected and can cause video distortion, making video motion magnification unsuitable for surgical guidance. The work in this thesis aims to address these shortcomings by excluding motion unrelated to the cardiovascular system from being exaggerated and enables motion magnification to become a viable image guidance option for detecting unseen vasculature. A higher order of motion temporal filter is proposed to negate the visible motion caused by respiration from the video motion magnification algorithm, whilst still remaining active on other present physiological motions. Tool motion is accounted for by creating a periodic filter prior to the insertion of surgical instruments into the scene. This filter acts as a comparative threshold and reduces the motion magnification effect where the tool motion is present. Both filters are demonstrated to work theoretically on synthetic motion profiles, as well as on a variety of surgical videos, where the Structural Similarity Image Matrix index is used to validate the filters’ performance. Spatio-temporal cross-sections of the videos also show an increase in usability. Finally, real-time system considerations are investigated with regard to temporal filter selection and suitability. Filter group delay, causal dependency, overall image quality and algorithmic complexity are considered a metrics for selecting the most suitable filter for a real-time usage case. Videos generated and used in this thesis are available from https://tinyurl.com/MotionMagForSurgicalVideo.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Motion Magnification for Surgical Video |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2022. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science UCL > Provost and Vice Provost Offices > UCL BEAMS UCL |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10152689 |
Archive Staff Only
 |
View Item |
