Add abstract
Want to add your dissertation abstract to this database? It only takes a minute!
Search abstract
Search for abstracts by subject, author or institution
Want to add your dissertation abstract to this database? It only takes a minute!
Search for abstracts by subject, author or institution
X-ray micro computed tomography (u-CT) image enhancement with artificial intelligence (AI)
by Antti Tilja
| Institution: | Tampere University |
|---|---|
| Department: | |
| Degree: | |
| Year: | 2022 |
| Keywords: | micro-computed tomography |
| Posted: | 3/25/2025 |
| Record ID: | 2222852 |
| Full text PDF: | https://trepo.tuni.fi/handle/10024/140449 |
Micro computed tomography and X-ray imaging as a whole are commonly used imaging techniques in research and medicine. As artificial intelligence has become more and more commonly used over the years it has also proven its place in medical imaging. Machine learning is used in most fields of research and the development of deep learning has improved the efficiency of AI even more. AI has already made detecting diseases and making diagnoses easier and faster and in the future, there will be even more improvements in the field of medical technology due to artificial intelligence. The purpose of this thesis was to examine how micro computed tomography images can be improved with artificial intelligence. Moreover, the image reconstructions were done with fewer projections than normally to see if image quality could be sustained with a lower projection count. Usually in CT imaging the projection count of reconstructions is around one thousand to acquire adequate image quality. The images used in this thesis were reconstructed with 1600, 800, 400, 200, and 100 projections. Dragonfly software was used for the AI. AI models were trained with different parameters to find out the best way to enhance image quality. As an overall result it can be said that with proper training data, AI is a very viable tool for enhancing images. The best results were gained with 800 and 400 projection images. The results show that the network training data should be similar to the image enhanced and filtering the output data also helps improve image quality. 200 and 100 projection image quality could not be improved enough to have practical use. The images taken with different exposure times had so few detectable differences that there was nothing that could be concluded from them.Mikro-tietokonetomografia, ja röntgenkuvantaminen ylipäänsä, ovat yleisesti käytettyjä kuvantamistapoja sekä tieteellisessä tutkimuksessa että lääketieteessä. Yksi suurimmista kehityssuunnista lääketieteellisessä kuvantamisessa nykyään on tekoälyn käyttö, joka on yleistynyt runsaasti viime vuosien aikana. Koneoppiminen on jo käytössä suurimmassa osassa tieteen aloista. Syväoppimisen kehitys ja suuremmat laskentatehot ovat mahdollistaneet entistä tehokkaampien tekoälyjen toteuttamisen. Tekoälyn avulla sairauksia pystytään jo havaitsemaan ja hoitamaan paremmin, ja tulevaisuudessa tekoälyllä tulee olemaan vielä suurempi rooli lääketieteen alalla ja diagnostiikassa. Tämän opinnäytetyön tarkoitus oli tutkia tekoälyn kykyä parantaa mikro tietokonetomografiakuvien laatua. Kuvat olivat rekonstruktioitu erilaisilla projektiomäärillä. Tämän tarkoitus oli tutkia, kuinka paljon pienemmillä projektiomäärillä rekonstruktioitujen kuvien laatua pystyy parantamaan. Yleensä tietokonetomografiassa käytetty projektiomäärä on noin 1000, jotta kuvanlaatu on riittävä. Tässä työssä käytetyt kuvat olivat rekonstruktioitu 1600, 800, 400, 200 ja 100 projektiolla. Tekoälyn opettamiseen käytettiin Dragonfly- tietokoneohjelmaa. Tekoälyjen opetuksessa käytettiin erilaisia neuroverkkorakenteita ja…
Want to add your dissertation abstract to this database? It only takes a minute!
Search for abstracts by subject, author or institution
|
|
Electric Cooperative Managers' Strategies to Enhan...
|
|
|
Bullied!
Coping with Workplace Bullying
|
|
|
The Filipina-South Floridian International Interne...
Agency, Culture, and Paradox
|
|
|
Commodification of Sexual Labor
Contribution of Internet Communities to Prostituti...
|
|
|
The Census of Warm Debris Disks in the Solar Neigh...
|
|
|
Solution or Stalemate?
Peace Process in Turkey, 2009-2013
|
|
|
Performance, Managerial Skill, and Factor Exposure...
|
|
|
The Deritualization of Death
Toward a Practical Theology of Caregiving for the ...
|
|
|
Emotional Intelligence and Leadership Styles
Exploring the Relationship between Emotional Intel...
|
|
|
Risk Factors and Business Models
Understanding the Five Forces of Entrepreneurial R...
|