Deep Burst Image Deblurring
Todorov, Peter (2020)
2020
Tietotekniikan DI-tutkinto-ohjelma - Degree Programme in Information Technology, MSc (Tech)
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2020-05-20
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202004143205
https://urn.fi/URN:NBN:fi:tuni-202004143205
Tiivistelmä
In the past two decades, mobile phone imaging has grown significantly. The camera is one of the main features of a new mobile phone and a lot of research is been done in this field to improve image quality. The camera module installed in mobile phones is restricted by the thin structure of the phones which means that no thick lenses or large image sensors are installed on the phones. That affects the amount of light that is captured which again affects the image quality. By having less light, phones perform worse in image quality measures compared to digital single-lens reflex (DSLR) cameras. Therefore, post-processing is applied to improve image quality.
Dark conditions and rapid motion are challenging situations in mobile imaging. In dark conditions, more light is gathered by having longer exposure time which can cause motion blur artifacts. Motion blur artifact can also be caused in daylight by a fast-moving object, for example, when photographing a race car. Motion blur causes the loss of sharp details and thus results in poor image quality. The method of removing blur from images and making them look sharper is called deblurring.
Recently, deep learning-based approaches have become popular in the field of signal processing. The results have been promising with these approaches because deep learning algorithms can learn and model nonlinear and complex relationships. In image restoration tasks, deep learning algorithms have also been applied to many tasks as also done here.
Dark conditions and rapid motion are challenging situations in mobile imaging. In dark conditions, more light is gathered by having longer exposure time which can cause motion blur artifacts. Motion blur artifact can also be caused in daylight by a fast-moving object, for example, when photographing a race car. Motion blur causes the loss of sharp details and thus results in poor image quality. The method of removing blur from images and making them look sharper is called deblurring.
Recently, deep learning-based approaches have become popular in the field of signal processing. The results have been promising with these approaches because deep learning algorithms can learn and model nonlinear and complex relationships. In image restoration tasks, deep learning algorithms have also been applied to many tasks as also done here.