Resolution-enhanced Digital Epiluminescence Microscopy Using Deep Computational Optics
| dc.contributor.author | Kabiljagic, Dino | |
| dc.date.accessioned | 2021-03-17T13:20:13Z | |
| dc.date.available | 2021-03-17T13:20:13Z | |
| dc.date.issued | 2021-03-17 | |
| dc.date.submitted | 2021-03-04 | |
| dc.description.abstract | Melanoma is the most common type of cancer, and the standard practice used for examining skin lesions is dermoscopy, where dermatologists use an epiluminescence microscope (ELM) to visualize the skin's surface and subsurface structures for anomalies. Conventional ELM instruments are being replaced by digital ELM instruments that enable dermatologists and other health care practitioners to digitally capture, archive, and analyze skin lesions using computer-aided diagnosis (CAD) software. One of the limiting factors of digital ELMs is a trade-off between spatial resolution and field of view (FOV), where a large FOV, which is needed to allow for larger skin lesions to be examined in their entirety, can be achieved by reducing magnification at the cost of spatial resolution (leading to a loss of fine details that can be indicative of malignancy and disease). In this thesis, we introduced the deep computation optics (DCO) framework for the purpose of resolution-enhanced digital ELM to improve the balance between spatial resolution and FOV. More specifically, the multitude of parameters of a deep computational model for numerically magnifying digital ELM images were learned through a wealth of low-resolution and high-resolution digital ELM image pairs. The proposed DCO approaches were experimentally validated, demonstrating improvements in the spatial resolution of the resolution-enhanced digital ELM when compared to more conventional methods, such as bicubic interpolation. Furthermore, we have demonstrated that the spatial resolution-enhancement improvements can be made within the deep computational models themselves where the model's receptive field is of the utmost importance since the missing information is better estimated when there is a larger number of neighbouring pixels involved. | en |
| dc.identifier.uri | http://hdl.handle.net/10012/16851 | |
| dc.language.iso | en | en |
| dc.pending | false | |
| dc.publisher | University of Waterloo | en |
| dc.subject | skin cancer | en |
| dc.subject | melanoma | en |
| dc.subject | epiluminescence microscopy | en |
| dc.subject | dermatology | en |
| dc.subject | deep computational optics | en |
| dc.title | Resolution-enhanced Digital Epiluminescence Microscopy Using Deep Computational Optics | en |
| dc.type | Master Thesis | en |
| uws-etd.degree | Master of Applied Science | en |
| uws-etd.degree.department | Systems Design Engineering | en |
| uws-etd.degree.discipline | System Design Engineering | en |
| uws-etd.degree.grantor | University of Waterloo | en |
| uws-etd.embargo.terms | 0 | en |
| uws.contributor.advisor | Wong, Alexander | |
| uws.contributor.affiliation1 | Faculty of Engineering | en |
| uws.peerReviewStatus | Unreviewed | en |
| uws.published.city | Waterloo | en |
| uws.published.country | Canada | en |
| uws.published.province | Ontario | en |
| uws.scholarLevel | Graduate | en |
| uws.typeOfResource | Text | en |