Syah, Maulidya Prastita (2026) Identifikasi Kredibilitas Sertifikat Prestasi Calon Mahasiswa Jalur SNBP dengan Metode Improved Scale Invariant Feature Transform dan Random Sample Consensus. Undergraduate thesis, UPN Veteran Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
22083010039-Cover.pdf Download (2MB) |
|
|
Text (Bab 1)
22083010039_bab 1.pdf Download (30kB) | Preview |
|
|
Text (Bab 2)
22083010039_bab 2.pdf Restricted to Repository staff only until 9 July 2029. Download (765kB) |
|
|
Text (Bab 3)
22083010039_bab 3.pdf Restricted to Repository staff only until 9 July 2029. Download (344kB) |
|
|
Text (Bab 4)
22083010039_bab 4.pdf Restricted to Repository staff only until 9 July 2029. Download (2MB) |
|
|
Text (Bab 5)
22083010039_bab 5.pdf Download (15kB) | Preview |
|
|
Text (Daftar Pustaka)
22083010039_daftar pustaka.pdf Download (177kB) | Preview |
|
|
Text (Lampiran)
22083010039_lampiran.pdf Restricted to Repository staff only until 9 July 2029. Download (332kB) |
Abstract
The National Selection Based on Achievement (SNBP) is an admission pathway for new students that takes into account students’ academic and non-academic achievements. Achievement certificates serve as supporting documents in this selection process. However, the large number of applicants and the wide variety of certificate formats and designs pose challenges for manual verification, such as the significant time required, the potential for misidentification, and the possibility of encountering unverifiable documents. Therefore, a system capable of automatically, quickly, and consistently identifying certificates is needed. This study aims to develop a certificate identification system using the Improved Scale Invariant Feature Transform (Improved SIFT) and Random Sample Consensus (RANSAC) methods. Improved SIFT is used to extract visual features such as logos and stamps, while RANSAC is used to filter out mismatched feature pairs. The study utilized 200 images of SNBP achievement certificates, which underwent the stages of data collection, labeling, preprocessing, feature extraction, feature matching, geometric validation, and evaluation using reprojection error and the structural similarity index measure. The system is also equipped with a feature to identify certificate levels and validate information via a QR code as supplementary information. The evaluation results yielded an average mean reprojection error of 28.47 pixels, an average inlier accuracy of 0.81 (81%) for certificates labeled as non-credible, and 8 certificates with SSIM values above 70%, indicating a high level of visual similarity to the reference images. Furthermore, all stages of the research were successfully implemented into a desktop application, enabling an integrated certificate identification process. The research results demonstrate that the combination of Improved SIFT and RANSAC is capable of automatically identifying visual elements on certificates and supports the document verification process in the SNBP selection.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > QA Mathematics > QA76.6 Computer Programming | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Data Science | ||||||||||||
| Depositing User: | Maulidya Prastita Syah | ||||||||||||
| Date Deposited: | 09 Jul 2026 04:02 | ||||||||||||
| Last Modified: | 09 Jul 2026 04:02 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/54618 |
Actions (login required)
 |
View Item |