PUTRA, YUSDIANSYA and Nurmaini, Siti (2022) PENINGKATAN KUALITAS CITRA JANTUNG JANIN MENGGUNAKAN PENDEKATAN SUPER RESOLUTION. Undergraduate thesis, Sriwijaya University.
![[thumbnail of RAMA_56201_09011381823078.pdf]](http://repository.unsri.ac.id/style/images/fileicons/text.png) |
Text
RAMA_56201_09011381823078.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy |
|
Text
RAMA_56201_09011381823078_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (8MB) | Request a copy |
Preview |
Text
RAMA_56201_09011381823078_0002085908_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (747kB) | Preview |
|
Text
RAMA_56201_09011381823078_0002085908_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (414kB) | Request a copy |
|
Text
RAMA_56201_09011381823078_0002085908_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (688kB) | Request a copy |
|
Text
RAMA_56201_09011381823078_0002085908_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy |
|
Text
RAMA_56201_09011381823078_0002085908_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (33kB) | Request a copy |
|
Text
RAMA_56201_09011381823078_0002085908_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (111kB) | Request a copy |
|
Text
RAMA_56201_09011381823078_0002085908_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (769kB) | Request a copy |
Abstract
The image of the fetal heart is the image produced by ultrasound, where the image is used in the health sector to obtain information about the development of the fetus in the womb. The image produced by ultrasound does not fully obtain detailed and detailed information. However, in the medical field, two-dimensional ultrasound, in general, is still widely applied to obstetricians because it is still considered appropriate in obstetrical examinations. This becomes one of the challenges, especially in the poor quality of the fetal heart image, both in the protocol and the different variations in each patient. Therefore, to overcome these problems, it is necessary to increase the resolution and accuracy of single image super-resolution which is faster on deep learning. In this research, image quality improvement uses a deep learning method with a super-resolution approach. The deep learning methods used are super-resolution residual network, super�resolution generative adversarial network, and super-resolution convolutional neural network. Image quality improvement is carried out by using low resolution and high-resolution images or enhancements from the initial video data, with a total of 144 models designed. The data will be trained and tested using fetal heart image data which are few in data access. Each model is designed with a combination of parameters such as epoch, batch size, learning rate, optimizer, and upscale. From the results of testing 144 models that have been designed, the model that produces the best performance is model 9 using the super-resolution convolutional neural network method, the model uses parameters epoch 1000, batch size 64, learning rate 0.0001, and Adam optimizer. This model produces the best evaluation with MSE 11.24343, SSIM 0.96637/96.637, and PSNR 34.76977 dB, while the unseen data results obtained are MSE 11.28212, SSIM 0.96082/96.082, and PSNR 34.96804 dB.
Actions (login required)
 |
View Item |