Efficient object detection method to Correcting images and videos from medium-resolution cameras using BEMD-based Scale-Invariant-Features-Transform: road safety application
DOI:
https://doi.org/10.22399/ijcesen.5263Keywords:
SIFT, Scale-Invariant-Features-Transform , BEMD, KLT algorithm , video trackingAbstract
Abstract:
The Scale-Invariant-Features-Transform (SIFT) is a computer vision algorithm to detect and match local features in images.
This algorithm is an indirect detection method since it doesn’t consider the entire candidate object to be detected but works only on the most interesting points of the object which allows efficiency in blurred images and Real Time applications such as object recognition, video tracking and many other applications.
The mean idea in the SIFT algorithm is to apply a derivative function in multiple scales. Then search the details (invariant key-points) across the derived scales.
To create the scales, firstly, octaves are made using different sub-samplings on the template image containing the object to be detected or matched. Then multiple pass-bands Gaussian filters are applied on each octave to have its Gaussian-filtered scales.
Then, the Laplacian filter is applied on each Gaussian-filtered scale to define scale-details called LOG (Laplacian of Gauss) on which the key-points will be searched. The invariant details called key-points will be the extrema points that persist as local extrema across the LOGs of the same octave scales.
In this paper, we do the same to define the different scales and we propose to apply an improved version of the BEMD (Bi-dimensional Empirical Mode Decomposition) on the obtained different Gaussian-filtered scales to define the details which are the Intrinsic Mode Functions IMFs.
In the proposed algorithm, since the IMFs are already obtained basing on the extrema points in a sifting process of the BEMD, no search will be done for the extrema points (key-points) across the BEMD-sifted components of Gaussian-filtered scales. In addition, since the BEMD gives locally ordered oscillating zero means components (the IMFs) from the most high spatial frequencies to the lowest ones then the continuous component (residue), the sifting process of the BEMD can be then stopped on only the first obtained IMF (containing the finest details) of each Gaussian-filtered scale which allows a computationally fast algorithm that can be used to quickly detect objects in a massive volume of data such as videos. Also, since an IMF is an oscillating mono-component free from any continuous component, the proposed BEMD-based SIFT algorithm allows correct detection of the object in different conditions such as blurred images, different sizes of the object and rotated objects.
In this paper, in a first step, the proposed BEMD-based SIFT algorithm is applied to recognise the prohibitory traffic signs especially for speed limitation.
In a second step, the speeding vehicles can be then automatically tracked by using the KLT algorithm (Kanade-Lucas-Tomasi) using motion vector thresholding in accordance with the already recognized sign of speed limitation by the proposed algorithm of detection.
References
[1] Jean Christophe Cexus." Analyse des signaux non-sationnaires par transformation de Huang, Opérateur de Teager-Kaizer, et transformation de Huang-teager THT ". Thesis 2005-University of Rennes1. .
[2] N.E. Huang, Z. Shen, S.R. Long, M.C. Wu, H.H. Shih, Q. Zheng, N.C. Yen, C.C.Tung et H.H. Liu. «The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis ". Proceedings of the Royal Society of London, 454:903–995, 1998.
[3] X.Li, J.Su & L.Yang,” Building Detection in SAR Images Based on Bi-Dimensional Empirical Mode Decomposition Algorithm” IEEE Geoscience and Remote Sensing Letters, (2019), 1–5. doi:10.1109/lgrs.2019.2928965.
[4] W.L.Hou, R.S.Jia, H.M.Sun, X.L.Zhang, M.D.Deng, & A.U. Tian, “Random Noise Reduction in Seismic Data by Using Bidimensional Empirical Mode Decomposition and Shearlet Transform.”, IEEE Access, 7, 71374–71386. (2019). doi:10.1109/access.2019.2920021.
[5] M.Chang, X.You & Cao.Z 2019. Bidimensional Empirical Mode Decomposition for SAR Image Feature Extraction with Application to Target Recognition. IEEE Access, 1–1. doi:10.1109/access.2019.2941397.
[6] Liu Yi-bing and Wu Qi, Ma Zhi-yong, Yan Ke-guo. «An Improved Hilbert-Huang Transform and Its Application in Faults Signal Analysis". Proceedings of the 2006 IEEE International Conference on Mechatronics and Automation. June 25-28, 2006, Luoyang, China. IEEE 2006.
[7] Ruqiang Yan. "Transient Signal Analysis Based on Hilbert-Huang Transform".IMTC 2005- Instrumentation and Mesurement. Technology Conference. Ottawa, Canada, 17-19 may 2005© IEEE 2005.
[8] Bin Shen. "Estimating the Instantaneous Frequencies of a multicomponent AM-FM Image by bidimensinal Empirical Mode Decomposition".1-3 September, 2005. Farco, Portugal.© 2005 IEEE.
[9] Lulu Hem Hongyuan Wang. "Spatial-variant Image Filtering Based on Bidimensional Empirical Mode Decomposition". The 18th International Conference on Pattern Recognition (ICPR'06). IEEE 2006.
[10] Wen-yang DUAN, Li-min HUANG, Yang HAN, Ya-hui ZHANG, Shuo HUANG, “A hybrid AR-EMD-SVR model for the short-term prediction of nonlinear and non-stationary ship motion”, J Zhejiang Univ-Sci A (Appl Phys& Eng) 2015 16(7):562-576.
[11] Xu Guanlei, Wang Xiaotong, Zhou Lijia, Xu Xiaogang, “Image decomposition and texture analysis via combined bi-dimensional Bedrosian's principles”, IET Image Process., 2018, Vol. 12 Iss. 2, pp. 262-273 © The Institution of Engineering and Technology 2017.
[12] Anna Linderhed, “Variable Sampling of the Empirical Mode Decomposition of Two-Dimensional signals”, International Journal of Wavelets, Multiresolution and Information Processing Vol. 3, No. 3 (2005) 435–452 © World Scientific Publishing Company.
[13] Anna Linderhed,” Adaptative image compression with wavelet packets and Empirical Mode Decomposition”. Thesis 2004, Institute of technology, Linköping University.
[14] M. Unser, “Spline, A Perfect Fit for Signal and Image Processing”, IEEE Signal Processing Magazine, Nov 1999, pp. 22-38.
[15] Paul G. Howard and Jeffrey Scott Vitter, Fellow, “Arithmetic Coding for Data Compression“, Proceedings of the IEEE. VOL. 82, NO. 6, JUNE 1994.
[16] Christophe Damerval, Sylvain Meignen, and Valérie Perrier, “A Fast Algorithm for Bidimensional EMD“, IEEE SIGNAL PROCESSING LETTERS, VOL. 12, NO. 10, OCTOBER 2005.
[17] F. Marvasti, “Non uniform sampling theorems for bandpass signals at or below the Nyquist density”, IEEE Transactions on Signal Processing, Vol. 44, No. 3, March 1996, pp. 572 -576.
[18] LINDERHED, A. (2009). “IMAGE EMPIRICAL MODE DECOMPOSITION: A NEW TOOL FOR IMAGE PROCESSING”. Advances in Adaptive Data Analysis, 01(02), 265–294. doi:10.1142/s1793536909000138
[19] Abdelouahed SABRI, Mohamed KAROUD, Hamid TAIRI and Aabdelah AARAB,” Image Watermarking Using the Empirical Mode Decomposition”, SETIT 2009 5th International Conference: Sciences of Electronic, Technologies of Information and Telecommunications March 22-26, 2009 – TUNISIA.
.
[20] Andrade, A. O., Nasuto, S., Kyberd, P., Sweeney-Reed, C. M., & Van Kanijn, F. R. (2006). EMG signal filtering based on Empirical Mode Decomposition. Biomedical Signal Processing and Control, 1(1), 44–55. doi:10.1016/j.bspc.2006.03.003
[21] Lu Gan, Long Zhou, and Xinge You, “A new GPR image de-noising method based on BEMD”, 2014 IEEE international conference on security, Pattern Analysis, and Cybernetics (SPAC), October 18-19, 2014, Wuhan, Hubei, China. ©2014 IEEE.
[22] JianJia Pan, YuanYan Tang, Dan Zhang, “A fractal-based BEMD method for image texture analysis”, 978-1-4244-6588-0/10/ ©201 0 IEEE.
[23] JIAN-JIA PAN, YUANYAN TANG , “A new image denoising method based on BEMD and SELF-SIMILAR FEATURE », Proceedings of the 2010 International Conference on Wavelet Analysis and Pattern Recognition, Qingdao, 11-14 July 2010, 978-1-4244-6531-6/10©2010 IEEE.
[24] Yun Qi, Jinkai Xu, Zhanjiang Yu and Huadong Yu, “The Reference Plane by BEMD for 3D Roughness Evaluation of Micro-milling Surface”, Proceedings of 2017 IEEE International Conference on Mechatronics and Automation August 6 - 9, Takamatsu, Japan. 978-1-5090-6759-6/17©2017 IEEE.
[25] Lowe, D. G. (1999). “Object recognition from local scale-invariant features”. Proceedings of the Seventh IEEE International Conference on Computer Vision. doi:10.1109/iccv.1999.790410
[26] D. Lowe, “ Distinctive Image Features from Scale-Invariant Keypoints “, International Journal of Computer Vision, vol. 60, 2004, p. 91 (DOI 10.1023/B:VISI.0000029664.99615.94.
[27] J. Canny, “ A Computational Approach To Edge Detection “, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 8, 1986, p. 679–714 (DOI 10.1109/TPAMI.1986.4767851).
[28] Abdelkader Moustafa Radwane Ghellab (2024), « Efficient image fusion method using improved Bi-dimensional Empirical Mode Decomposition », International Journal of Image and Data Fusion, 15:1, 44-72, DOI: 10.1080/19479832.2024.2307605
[29] Simon Baker, Iain Matthews, "Lucas-Kanade 20 Years On: A Unified Framework". International Journal of Computer Vision, February 2004, Volume 56, issue 3, pp 221-255.
[30] Rosiani, U. D., Choirina, P., & Shoumi, M. N. (2021). Micro-expression recognition based on motion detection method. IOP Conference Series: Materials Science and Engineering, 1073(1), 012069. doi:10.1088/1757-899x/1073/1/012069
[31] Piao J C and Kim S D 2017 Adaptive monocular visual–inertial SLAM for real-time augmented reality applications in mobile devices Sensors 17(11) 2567
[32] Binwei Weng, Manuel Blanco-Velasco, and Kenneth E.Barner 2006. ECG Denoising Based on the Empirical Mode Decomposition. International Conference of the 28th IEEE Engineering in Medicine and Biology Society. doi:10.1109/iembs.2006.259340. New York City, USA, Aug 30-Sept 3, 2006.
[33] A.T.Celebi & S.Erturk 2010. Empirical mode decomposition based visual enhancement of underwater images. 2nd International Conference on Image Processing Theory, Tools and Applications (2010). doi:10.1109/ipta.2010.5586758.
[34] JianJia Pan, YuanYan Tang and Dan Zhang 2010. A fractal-based BEMD method for image texture analysis. IEEE International Conference on Systems, Man and Cybernetics (2010). doi:10.1109/icsmc.2010.5642020.
[35] Z.He, Q.Wang, Y.Shen, J.Jin& Y.Wang 2013. Multivariate Gray Model-Based BEMD for Hyperspectral Image Classification. IEEE Transactions on Instrumentation and Measurement, 62(5), 889–904, (2013). doi:10.1109/tim.2013.2246917.
[36] Donghoh Kim, Minjeong Park, & Hee-Seok Oh 2012. Bidimensional Statistical Empirical Mode Decomposition. IEEE Signal Processing Letters(2012), 19(4), 191–194. doi:10.1109/lsp.2012.2186566.
[37] Zhongxuan Liu & Peng.S 2005. Boundary Processing of bidimensional EMD using texture synthesis. IEEE Signal Processing Letters (2005), 12(1), 33– 36. doi:10.1109/lsp.2004.839700.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 International Journal of Computational and Experimental Science and Engineering
This work is licensed under a Creative Commons Attribution 4.0 International License.
