Deep learning denoising for tomographic reconstructions
Başlık çevirisi mevcut değil.
- Tez No: 841803
- Danışmanlar: PROF. MARCO MARCON
- Tez Türü: Yüksek Lisans
- Konular: İletişim Bilimleri, Communication Sciences
- Anahtar Kelimeler: Computed Tomography, Sinogram, Deep Learning, U-Net, CGAN, Denoising
- Yıl: 2023
- Dil: İngilizce
- Üniversite: Polıtecnıco Dı Mılano
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
Computed Tomography (CT) is a non-destructive imaging technique that can produce detailed internal reconstructions of objects and live beings in the form of 2D images or 3D models. As a result of the tomography process, sinograms are obtained and these are the equivalents of the scanned object in another domain. Many methods have been used so far to obtain real images from sinogram images. Even the most advanced methods aim to approximate the original image rather than directly obtaining it. Current studies also direct the images obtained to be closer to reality. Besides, reconstructed images may have some artifacts. Various denoising methods are used to remove these artifacts. Most methods apply the denoising process after the reconstruction process. In this thesis, the denoising process will be applied on the sinogram domain before reconstruction. As denoising methods, deep learning methods, which are increasingly popular today, will be used. U-Net, which shows high performance, especially in image processing applications, and Conditional GAN models, which have come to the fore with surprising results in recent years, will be used for denoising. These methods will be tested in test environments prepared after the necessary training processes, their outputs will be taken and necessary comparisons will be made and presented in an explanatory way. In addition, the preparation and application of the data set required for the training and testing of these models will be explained in this study. In addition to the synthetic data used, an evaluation is made on a real CT acquisition. This thesis aims to explore the potential of deep learning methods for denoising CT sinograms before reconstruction. The results of this study could have significant implications for the field of imaging, as it could improve the accuracy of CT imaging and lead to better CT observations.
Özet (Çeviri)
Özet çevirisi mevcut değil.
Benzer Tezler
- Pet görüntülerinde gürültü azaltıcı difüzyon olasılıksal modeller (DDPMs) ile gürültü azaltma
Noise reduction in pet images using denoising diffusion probabilistic models (DDPMs)
EMİNE PACCİ
Yüksek Lisans
Türkçe
2026
İstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
Yükleniyor...
- Beyin BT görüntülerinden inme tespiti ve sınıflandırılması için gürültü azaltma yöntemleri
Noise reduction methods for the stroke detection and classification from brain computed tomography images
HAKAN SÖKÜN
- Medikal görüntü analizinde gürültü saldırılarına karşı derin öğrenme modellerinin performanslarının karşılaştırılması
Benchmarking of deep learning models against adversarial attacks in medical image analysis
GÖKÇE OK
- A comparative analysis of deep learning models for multi-horizon cryptocurrency return forecasting with wavelet denoising
Dalgacık gürültü giderme ile kripto para getirilerinin çok ufuklu tahmininde derin öğrenme modellerinin karşılaştırmalı analizi
MEHMET ÖZSOY
- Deep learning for ınverse problems in ımaging
Görüntüleme ters problemlerinde derin öğrenme
HASAN HÜSEYİN KARAOĞLU
Doktora
İngilizce
2024
İstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
Yükleniyor...