Cascaded Architecture for Segmenting Prostate Cancer Lesions in Biparametric MRI
Full Metadata
| Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Sitthichok Chaichulee | - |
| dc.contributor.author | Chaloemphon Thipkasorn | - |
| dc.contributor.department | ?????????????? | - |
| dc.contributor.department | Graduate School | - |
| dc.date.accessioned | 2024-07-20 07:43 | - |
| dc.date.accessioned | 2026-02-11T02:42:21Z | - |
| dc.date.available | 2024-07-20 07:43 | - |
| dc.date.issued | 2024 | - |
| dc.description | ????????,??????????????,2567 | - |
| dc.description.abstract | Prostate cancer (PCa) is one of the most common cancers affecting men worldwide. Crucial is the early detection and appropriate treatment. Currently, PSA (prostate-specific antigen) tests are used by doctors to indicate cancerous lesions, and Magnetic Resonance Imaging (MRI) techniques are employed to identify affected areas. However, challenging is the task of reading MRI images to find the exact location of prostate lesions due to their complexity and variability. In this study, Convolutional Neural Networks (CNNs), a type of Deep Learning (DL) suited for image and video data, were used for image segmentation. A two-step approach was taken. First, the image was cropped by the SegResNet model to reduce the area needing segmentation for the gland. This step helps to focus after analysis on the region most likely to have cancer lessions. Second, another SegResNet model was used to segment the cancerous tumors within the cropped prostate area from the first step. Bi-parametric MRI images, including T2-Weighted (T2W), Apparent Diffusion Coefficient (ADC), and Diffusion- Weighted Imaging (DWI) sequences, were used. Training and validation of the model were done using the PI-CAI 2022 dataset. Additionally, external datasets such as Prostate-158 and PSU-mpMRI were used to test the model's performance. Reported were segmentation accuracies of 0.4211 for Prostate-158 and 0.3588 for PSU-mpMRI, with detected position distances of 1.8252 for Prostate-158 and 3.601 for PSU-mpMRI. | - |
| dc.description.abstract | ???????3????????? (PCa) ??U?????????????????????????????????????????3???'?????????? ??? ?????????????????????????'?????????'????????????????????U??????????? ???]??????????H???????'????????? ????????3? PSA ??????3??????????????????????????????????'??????????????????????3????????c???????? Magnetic Resonance Imaging (MRI) ??????U????????????????????????????????? ???????????????3????????? ??3 ????3????? MRI ???????????????3?????3????????????????????????????????????'??????????????????? ??? ?????????????'??'?????????? Convolutional Neural Networks (CNNs) ??????U????????????????? Deep Learning (DL) ???????????????????????????????'??????????????????????????????????3??3???????? (Segmentation) ???????????????? ??'??'???????????????????? ??????????????????????????????'????? SegResNet ??????????'?????3???????????3?????????? ?????????????????????3??3?????????3????????? ??????? ????3????'???????????H?????3?????'??????????????????U?????'????????????????????????????????? ??????????????????????? SegResNet ???????????'???????3??3???????????????????????????????????3???????????????????????? ????? ??????????????????'?????????????????? (bi-parametric MRI) ??'??3 T2-Weighted (T2W) ??? Apparent Diffusion Coefficient (ADC) ?????? Diffusion-Weighted Imaging (DWI) ??????????'????'???? PI-CAI 2022 ??????�??? (Training) ????????????????????????????? (Validation) ?????????????'????????????3?? Prostate-158 ??? PSU-mpMRI ??'??????????'??????????????????????????? (Testing) ???????????3?????? ?????3??3????????3??? 0.4211 ????????'???? Prostate-158 ??? 0.3588 ????????'???? PSU-mpMRI ????3? ?????3???????????3?????????????'???3??? 1.8252 ????????'???? Prostaet-158 ????'???? PSU-mpMRI ???3??? 3.601 | - |
| dc.identifier.uri | https://kb.psu.ac.th/handle/2025/19788 | - |
| dc.language.iso | en | - |
| dc.publisher | Prince of Songkla University | - |
| dc.rights | Attribution-NonCommercial-NoDerivs 3.0 Thailand | - |
| dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/3.0/th/ | - |
| dc.subject | Prostate Cancer | - |
| dc.subject | Image Segmentations | - |
| dc.subject | Bi-parametric Magnetic Resonance Imaging (bpMRI) | - |
| dc.subject | Multi-parametric Magnetic Resonance Imaging (mpMRI) | - |
| dc.title | Cascaded Architecture for Segmenting Prostate Cancer Lesions in Biparametric MRI | - |
| dc.title.alternative | Cascaded Architecture for Segmenting Prostate Cancer Lesions in Biparametric MRI | - |
| dc.type | Thesis | - |
Files
Files
Collections


