Analisis Trade-Off Akurasi dan Efisiensi Pada Algoritma Feature Matching Pada Rambu Lalu Lintas
- Malfino Wildan Akhya Universitas Dian Nuswantoro, Semarang, Indonesia
- Burhanuddin Dhika Universitas Dian Nuswantoro, Semarang, Indonesia
- Albireo Musyaffa Finoe Universitas Dian Nuswantoro, Semarang, Indonesia
-
Muhammad Naufal
*
Universitas Dian Nuswantoro, Semarang,
Indonesia
- Ricardus Anggi Pramunendar Universitas Dian Nuswantoro, Semarang, Indonesia (*) Corresponding Author
Abstract
The problem of visual feature detection and matching is becoming increasingly crucial with the development of Intelligent Transportation Systems (ITS), particularly in autonomous vehicles and ADAS, where the reliability of traffic sign recognition is highly affected by variations in lighting, rotation, and image scale. Therefore, a comprehensive comparison is needed to identify the algorithm that offers the best trade-off between accuracy and efficiency. This study compares three feature algorithms Scale Invariant Feature Transform (SIFT), KAZE, and Oriented FAST and Rotated BRIEF (ORB) on 52 Traffic Sign Image Pairs with a minimalist pre-processing pipeline. The algorithms' performance is evaluated based on Inlier Ratio (Accuracy), Total Execution Time (tTotal), and Keypoint Difference Ratio (Stability). The experimental results show that ORB is the most optimal choice as it excels in both crucial metrics, Efficiency 0.0056 seconds and Inlier Ratio Accuracy 0.6575. Meanwhile, KAZE shows the highest Detection Stability with Difference Ratio 23.27%, although hampered by the longest computation time. These findings recommend ORB for real-time systems and suggest the development of a hybrid pipeline that leverages the stability of KAZE and compares favorably with modern deep learning models
Downloads
Bila bermanfaat silahkan share artikel ini
Berikan Komentar Anda terhadap artikel Analisis Trade-Off Akurasi dan Efisiensi Pada Algoritma Feature Matching Pada Rambu Lalu Lintas
Pages: 105-113
Copyright (c) 2025 Malfino Wildan Akhya
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
