Search

In interior computed tomography (CT), the x-ray beam is collimated to a limited field-of-view (FOV) (e.g. the volume of the heart) to decrease exposure to adjacent organs, but the resulting image has a severe truncation artifact when reconstructed with traditional filtered back-projection (FBP) type algorithms. In some examinations, such as cardiac or dentomaxillofacial imaging, interior CT could be used to achieve further dose reductions. In this work, we describe a deep learning (DL) method to obtain artifact-free images from interior CT angiography. Our method employs the Pix2Pix generative adversarial network (GAN) in a two-stage process: (1) An extended sinogram is computed from a truncated sinogram with one GAN model, and (2) the FBP reconstruction obtained from that extended sinogram is used as an input to another GAN model that improves the quality of the interior reconstruction. Our double GAN (DGAN) model was trained with 10 000 truncated sinograms simulated from real computed tomography angiography slice images. Truncated sinograms (input) were used with original slice images (target) in training to yield an improved reconstruction (output). DGAN performance was compared with the adaptive de-truncation method, total variation regularization, and two reference DL methods: FBPConvNet, and U-Net-based sinogram extension (ES-UNet). Our DGAN method and ES-UNet yielded the best root-mean-squared error (RMSE) (0.03 +/- 0.01), and structural similarity index (SSIM) (0.92 +/- 0.02) values, and reference DL methods also yielded good results. Furthermore, we performed an extended FOV analysis by increasing the reconstruction area by 10% and 20%. In both cases, the DGAN approach yielded best results at RMSE (0.03 +/- 0.01 and 0.04 +/- 0.01 for the 10% and 20% cases, respectively), peak signal-to-noise ratio (PSNR) (30.5 +/- 2.6 dB and 28.6 +/- 2.6 dB), and SSIM (0.90 +/- 0.02 and 0.87 +/- 0.02). In conclusion, our method was able to not only reconstruct the interior region with improved image quality, but also extend the reconstructed FOV by 20%.

Development of the rubber compounds is started commonly in laboratory scale. Bringing the compounding process to the production scale requires successful scale-up. Various scale-up approaches have been suggested in the literature, but the studies have mainly focused on carbon black reinforced rubber compounds. This thesis covers a study of effect of time-temperature-and power peak-based scale-up on the properties of a silica reinforced compound. Initial small- and large-scale mixing programs were compared. Mixing programs differed from feeding orders, temperature profiles, total amount of revolutions, and mixing times. The biggest differences were at the first stage of the mixing programs. Differences in the mixing programs resulted in differences in Mooney viscosities, mechanical properties, and dynamic mechanical properties. Mixing programs were unified to study the effect of the mixer size on the compound properties. Unifications were done first by unifying the feeding orders of the mixers and second by unifying the temperature profiles. Time-temperature based upscaling was studied in three different mixer sizes: small (1.5 l), medium (45 l), and large (320 l). Mixing programs of the small- and medium-scale mixers were modified the way that their time-temperature profiles corresponded to the time-temperature profile of the large-scale mixer. Targeted time-temperature profile was obtained by increasing the TCU temperatures and rotor speeds along the decreasing mixer size. Mixing program changes increased the total amount of revolutions and silanisation time. This was reflected to the compound properties. Mooney viscosity, ΔG*, and dynamic mechanical properties decreased by decreasing mixer size. Effect of the scale-up on the mechanical properties was not that clear. Anyhow, it was seen that Mod10 was significantly higher in case of the large-scale mixer, whereas Mod300/Mod100 was higher and elongation at break lower in the case of the small-scale mixer. Results suggested that the increase in the total revolutions and in silanisation times resulted in better dispersion and higher degree of silanisation, which furthermore caused differences in the properties of the compounds. Power peak-based upscaling was studied as an alternative to time-temperature-based upscaling. Mixing programs used in the time-temperature-based upscaling were modified so that the mixing time after passing the last power peak was constant in all the mixer sizes. Consequently, the mixing program was terminated the earlier the smaller was the mixer. The total amount of revolutions and silanisation times were more uniform between different mixer sizes than in the time-temperature based upscaling. Changes in the total amount of revolutions and silanisation times reflected to the compound properties. Mooney viscosity, ΔG*, dispersion, dynamic mechanical properties, and Mod300/Mod100 were more uniform than when the mixing process was upscaled based on the time-temperature profile. Power peak-based scale-up was seen as better option for mixing program scale-up than time-temperature-based scale-up.