Perbandingan Performa SIFT dan ORB dalam Pengolahan Dataset Wajah nist_2
Abstract
The problem of detecting and matching facial features in digital images is becoming increasingly crucial with the development of biometrics and human computer interaction applications, especially under varying lighting, orientation, and expression conditions. In this study, two feature detection and matching algorithms Scale Invariant Feature Transform (SIFT) and Oriented FAST and Rotated BRIEF (ORB) are compared on grayscale images processed using adaptive local contrast and Gaussian filtering. The performance of both algorithms is quantitatively evaluated based on the number of keypoints, matching precision, execution time, and visual accuracy. Experimental results show that ORB has an execution time of about 4.7-fold faster than SIFT, indicating ORB’s suitability for real time applications. In contrast, SIFT produces a higher matching rate and shows better robustness to lighting variations and facial deformation. These findings provide practical guidelines for selecting algorithms based on priority: speed in real time applications or accuracy in challenging environmental conditions.
Kata kunci: Feature matching; Facial image processing; Adaptive local contrast; Scale Invariant Feature Transform; Oriented FAST and Rotated BRIEF.
Abstrak
Permasalahan deteksi dan pencocokan fitur wajah pada citra digital menjadi semakin krusial seiring berkembangnya aplikasi biometrik dan interaksi manusia komputer, terutama dalam kondisi pencahayaan, orientasi, dan ekspresi yang bervariasi. Dalam studi ini, dua algoritma deteksi dan pencocokan fitur Scale Invariant Feature Transform (SIFT) dan Oriented FAST and Rotated BRIEF (ORB) dibandingkan pada citra grayscale yang telah diproses menggunakan adaptive local contrast dan Gaussian filtering. Kinerja kedua algoritma dievaluasi secara kuantitatif berdasarkan jumlah titik kunci, presisi pencocokan, waktu eksekusi, dan akurasi visual. Hasil eksperimen menunjukkan bahwa ORB memiliki waktu eksekusi sekitar 4,7 lipat lebih cepat daripada SIFT, menandakan kecocokan ORB untuk aplikasi real time. Sebaliknya, SIFT menghasilkan tingkat kecocokan yang lebih tinggi dan menunjukkan ketahanan yang lebih baik terhadap variasi pencahayaan dan deformasi wajah. Temuan ini memberikan pedoman praktis dalam memilih algoritma sesuai prioritas: kecepatan pada aplikasi real time atau ketelitian dalam kondisi lingkungan yang menantang.
Keywords
References
G. Guo and N. Zhang, “A survey on deep learning based face recognition,” Computer Vision and Image Understanding, vol. 189, p. 102805, 2019. doi: 10.1016/j.cviu.2019.102805
B. Jafrasteh, S. P. Lubián-López, and I. Benavente-Fernández, “A deep sift convolutional neural networks for total brain volume estimation from 3D ultrasound images,” Computer Methods and Programs in Biomedicine, vol. 237, p. 107805, 2023. doi: 10.1016/j.cmpb.2023.107805
L. Cai, Y. Ye, X. Gao, Z. Li, and C. Zhang, “An improved visual SLAM based on affine transformation for ORB feature extraction,” Optik, vol. 223, p. 165421, 2020. doi: 10.1016/j.ijleo.2020.165421
L. Marlinda, S. Rustad, R.S. Basuki, F. Budiman, dan M. Fatchan, “Matching images on the face of a Buddha statue using the scale invariant feature transform (SIFT) method,” in Proc. 2020 7th International Conference on Information Technology, Computer, and Electrical Engineering (ICITACEE), Semarang, Indonesia, 24–25 Sept. 2020, pp. 169-172, doi: 10.1109/ICITACEE50144.2020.9239221.
Y. Xie, W. Tian, H. Zhang, and T. Ma, “Facial expression recognition through multi-level features extraction and fusion,” Soft Computing, vol. 27, pp. 11243–11258, 2023. doi: 10.1007/s00500-023-08531-z
K.C. Paul and S. Aslan, “An improved real-time face recognition system at low resolution based on local binary pattern histogram algorithm and CLAHE,” Optics and Photonics Journal, vol. 11, no. 4, pp. 63–78, 2021. doi: 10.4236/opj.2021.114005
G. Sanil, K. Prakasha K., S. Prabhu, V. Nayak, and A. Jayakala, “Region-wise landmarks-based feature extraction employing SIFT, SURF, and ORB feature descriptors to recognize Monozygotic twins from 2D/3D facial images,” F1000Research, vol. 14, art. no. 444, 2025. doi: 10.12688/f1000research.162911.1
M. Kennerley, “A comparison of SIFT, SURF and ORB on OpenCV,” Medium, 2021. [Online]. Available: A Comparison of SIFT, SURF and ORB on OpenCV | by Mikhail Kennerley | Medium. [Accessed: May 16, 2025].
I. Suárez, G. Sfeir, J.M. Buenaposada, and L. Baumela, “BEBLID: Boosted efficient binary local image descriptor,” Pattern Recognition Letters, vol. 133, pp. 366–372, May 2020. doi: 10.1016/j.patrec.2020.04.005
L. Marlinda, F. Budiman, R. Basuki, and A.Z. Fanani, “Comparison of SIFT and ORB methods in identifying the face of Buddha statue,” JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer), vol. 8, no. 2, pp. 139–144, Feb. 2023. doi: 10.33480/jitk.v8i2.4086
M. Isik, “Comprehensive empirical evaluation of feature extractors in computer vision,” PeerJ Computer Science, vol. 10, art. no. e2415, Nov. 2024. doi: 10.7717/peerj-cs.2415
K. Matusiak, P. Skulimowski, and P. Strumillo, “Unbiased evaluation of keypoint detectors with respect to rotation invariance,” IET Computer Vision, vol. 11, no. 7, pp. 547–555, Sep. 2017. doi: 10.1049/iet-cvi.2016.0434
R. Cui, M. Wen, K. Zhang, and C. Sun, “Contrast threshold adaptive adjustment algorithm for remote sensing image matching,” J. Appl. Remote Sens., vol. 15, no. 3, p. 036519, Sep. 2021, doi: 10.1117/1.JRS.15.036519.
O. Sidhom, H. Ghazouani, and W. Barhoumi, “Three-phases hybrid feature selection for facial expression recognition,” J. Supercomput., vol. 80, pp. 8094–8128, 2024, doi: 10.1007/s11227-023-05758-3.
N. Kaur, S. Rani, and S. Kaur, “Real-time video surveillance based human fall detection system using hybrid Haar cascade classifier,” Multimed. Tools Appl., vol. 83, pp. 71599–71617, 2024, doi: 10.1007/s11042-024-18305-w.
S. Xu, S. Chen, R. Xu, C. Wang, P. Lu, and L. Guo, “Local feature matching using deep learning: A survey,” Information Fusion, vol. 101, p. 102344, 2024, doi: 10.1016/j.inffus.2024.102344.
D. DeTone, T. Malisiewicz, and A. Rabinovich, “SuperPoint: Self-Supervised Interest Point Detection and Description,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. Workshops (CVPRW), Salt Lake City, UT, USA, Jun. 2018, pp. 224–236.
M. Dusmanu, I. Rocco, T. Pajdla, M. Pollefeys, J. Sivic, A. Torii, and T. Sattler, “D2-Net: A Trainable CNN for Joint Description and Detection of Local Features,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Long Beach, CA, USA, Jun. 2019, pp. 8096–8105.
J. Revaud, C. De Souza, M. Humenberger, and P. Weinzaepfel, “R2D2: Repeatable and Reliable Detector and Descriptor,” in Advances in Neural Information Processing Systems, vol. 32, NeurIPS 2019, Vancouver, Canada, Dec. 2019.
Z. Luo, L. Zhou, X. Bai, H. Chen, J. Zhang, Y. Yao, S. Li, T. Fang, and L. Quan, “ASLFeat: Learning Local Features of Accurate Shape and Localization,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Seattle, WA, USA, Jun. 2020, pp. 6589–6598.
H. Abdellali, R. Frohlich, V. Vilagos, and Z. Kato, “L2D2: Learnable Line Detector and Descriptor,” in Proc. 2021 Int. Conf. on 3D Vision (3DV), London, United Kingdom, Dec. 2021, doi: 10.1109/3DV53792.2021.00054.
R. A. Pramunendar, nist_2 Facial Image Dataset, “Grayscale facial images with varying lighting, orientation, and expression.” [Online]. Available: https://drive.google.com/drive/ folders/1xIomu91x1H0OJKqqWmrXlHuixvHRxumo [Accessed: Apr. 27, 2025].
How To Cite This :
Refbacks
- There are currently no refbacks.