Skripsi
PENERAPAN LIBRARY PYTORCH3D PADA REKONTRUKSI JARINGAN TRANSVENTRIKULAR CITRA ULTRASONOGRAFI DARI 2 DIMENSI KE 3 DIMENSI.
The reconstruction of two-dimensional images into three-dimensional using transventricular objects plays a crucial role in the medical field, particularly in the diagnosis and treatment within neurology. This research employs the Pytorch3D method to generate three-dimensional representations of transventricular image objects. The study begins by performing the segmentation process on the transventricular objects in an ultrasound dataset that has been verified and labeled by medical professionals. The segmentation process utilizes the U-Net architecture, which effectively performs the segmentation process. After selecting the data for segmentation, the training process can be carried out. The U-Net architecture demonstrates effective segmentation results. With the segmented output from U- Net, the study proceeds to the next step, which is the reconstruction process. The reconstruction process transforms the two-dimensional images into three- dimensional representations. This is accomplished by utilizing two models to enhance the results. In previous research, the PIFUHD architecture was used for this purpose. The PIFUHD method can generate three-dimensional representations with a single input, which is the segmented image output from the previous step. The resulting representation is then implemented into the Pytorch3D method. The objects generated through this process will be similar to the ones produced by the PIFUHD method, which are three-dimensional objects. This research has significant implications in the medical field and contributes to the advancement of neurology-related diagnosis and treatment
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2307006463 | T130964 | T1309642023 | Central Library (Referens) | Available |
No other version available