Akçaağaç türlerinin taramalı elektron mikroskobu görüntüleri kullanılarak otomatik olarak sınıflandırılması

Öz

Ahşap, yaylı çalgı yapımında kilit bir role sahiptir. Yaylı çalgılar genellikle bu konuda baskın olan Akçaağaç türlerinden yapılır. Ahşap türlerinin doğru sınıflandırılması, yaylı çalgıların sahtekarlık olmadan yüksek kaliteli malzemeler kullanılarak yapılması için çok önemlidir. Bu çalışmada, üç farklı Akçaağaç türüne ait altı farklı sınıfın taramalı elektron mikroskobu (SEM) görüntülerini doğru bir şekilde sınıflandırmak için akıllı bir uygulama geliştirilmiştir. Her bir sınıfa ait SEM görüntüleri ayrı ayrı farklı boyutlarda altı alt bölgeye ayrılmıştır. Her alt bölgede 11 özellik çıkarılmış ve her sınıf için sayısal veri kümeleri oluşturulmuştur. Çıkarılan özelliklerin etkinliği için tek değişkenli seçim, özellik önemi ve ısı haritası ile korelasyon matrisi olmak üzere üç özellik seçim tekniği uygulanmıştır. Akçaağaç türlerinin SEM görüntüleri, doğrudan sınıflandırma ve ikili sınıflandırma olmak üzere iki farklı yaklaşıma dayalı olarak beş kat çapraz doğrulama altında makine öğrenimi modelleri tarafından sınıflandırılmıştır. Doğrudan sınıflandırma yaklaşımına dayalı en iyi makine öğrenmesi modeli %82.3 doğruluk oranıyla Kuadratik DVM modeli olarak belirlenmiştir. İkili sınıflandırma yaklaşımının genel doğruluğu ise Kuadratic DVM ve Ensemble subspace discriminant (ESD) modellerinin birlikte çalışması sonucunda %92.1 olarak hesaplanmıştır. Bu çalışma temel olarak Akçaağaç türlerine ait SEM görüntülerinin sınıflandırılması, alt bölge analizi, özellik çıkarımı ve seçimi ile makine öğrenmesi modellerinin karşılaştırılmasına odaklanmaktadır.

Anahtar Kelimeler

• Akçaağacı
• Sınıflandırma
• Makine Öğrenmesi
• SEM Görüntüleri

Automated classification of wood types of acer using scanning electron microscopy images

Öz

Wood has a key role for string instrument making. String instruments are generally made of wood types of Acer which is dominant for this issue. Accurate classification of wood types is pivotal that string instruments must be made by using high qualified materials without fraud. In this work, an innovative application was implemented to accurately classify scanning electron microscopy (SEM) images of the six different classes belonging to three different wood types of Acer. SEM images of each class were individually divided into six subregions of different sizes. 11 features were extracted on each subregion, thus creating the numerical datasets for each class. For the effectiveness of the extracted features, three feature selection techniques, namely univariate selection, feature importance and correlation matrix with heatmap were applied. SEM images of wood types of Acer were classified by machine learning (ML) models under five-fold cross validation based on two different approaches as direct classification and binary classification. The best ML model based on direct classification approach was determined as Quadratic Support Vector Machine (SVM) model with accuracy of 82.3%. General accuracy of the binary classification approach was calculated as 92.1% as a result of the collaboration of Quadratic SVM and Ensemble subspace discriminant (ESD) models. This study mainly focuses on classification of SEM images of wood types of Acer, subregion analysis, feature extraction and selection, and comparison of ML models.

Anahtar Kelimeler

• Acer
• Classification
• Machine Learning
• SEM Images

Kaynakça

1. [1] Yaygingol HS. Yaylı Çalgı Yapım Teknolojisi, 3. Baskı, Eskişehir, Türkiye, Anadolu Üniversitesi Yayınları, 2010.
2. [2] Nicolini G, Scolari G. Come Nasce un Violin, Cremona, Italy, Edizioni Stradivari, 1985.
3. [3] Gökmen H. Kapalı Tohumlular (Angiospermae), 2. Baskı. Ankara, Türkiye, Orman Bakanlığı Orman Genel Müdürlüğü, 1977.
4. [4] Salma Gunawan PH, Prakasa E, Sugiarto B, Wardoyo R, Rianto Y, Damayanti R, Krisdianto Dewi LM. “Wood identification on microscopic image with daubechies wavelet method and local binary pattern”. International Conference on Computer, Control, Informatics and its Applications, Tangerang, Indonesia, 1-2 November 2018.
5. [5] Zamri MIP, Khairuddin ASM, Mokhtar N, Yusof R. “Wood species recognition system based on improved basic grey level aura matrix as feature extractor”. Journal of Robotics, Networking and Artificial Life, 3(3), 140-143, 2016.
6. [6] Filho PLP, Oliveira LS, Nisgoski S, Britto Jr. AS. “Forest species recognition using macroscopic images”. Machine Vision and Applications, 25, 1019–1031, 2014.
7. [7] Yusof R, Khalid M, Khairuddin ASM. “Application of kernel-genetic algorithm as nonlinear feature selection in tropical wood species recognition system”. Computers and Electronics in Agriculture, 93 68–77, 2013.
8. [8] Mohamed A, Abdullah A. “Scanning electron microscopy (SEM): a review”. 2018 International Conference on Hydraulics and Pneumatics–HERVEX, Baile Govora, Romania, 7-9 November 2018.

9. [9] Pertuz S, Puig D, Garcia MA. “Analysis of focus measure operators in shape-from-focus”. Pattern Recognition, 46(5), 1415–1432, 2012.
10. [10] Pech-Pacheco J, Cristobal G, Chamorro-Martinez J, Fernandez-Valdivia J. “Diatom autofocusing in bright field microscopy: a comparative study”. 15th International Conference on Pattern Recognition, Barcelona, Spain, 3-7 September 2000.
11. [11] Kavsaoğlu AR, Sehirli E. “A novel study to classify breath inhalation and breath exhalation using audio signals from heart and trachea”. Biomedical Signal Processing and Control, 80, 1-9, 2023.
12. [12] Firestone L, Cook K, Culp K, Talsania N, Preston Jr. K. “Comparison of autofocus methods for automated microscopy”. Cytometry, 12, 195–206, 1991.
13. [13] Huang W, Jing Z. “Evaluation of focus measures in multi-focus image fusion”. Pattern Recognition Letters, 28(4), 493–500, 2007.
14. [14] Malik AS, Choi TS. “A novel algorithm for estimation of depth map using image focus for 3D shape recovery in the presence of noise”. Pattern Recognition, 41(7), 2200–2225, 2008.
15. [15] Geusebroek J, Cornelissen F, Smeilders A, Geerts H. “Robust auto focusing in microscopy”. Cytometry, 39, 1–9, 2000.
16. [16] Santos A, de Solorzano CO, Vaquero JJ, Pena JM, Mapica N, Pozo FD. “Evaluation of auto focus functions in molecular cytogenetic analysis”. Journal of Microscopy, 188(3), 264–272, 1997.
17. [17] Sun Y, Duthaler S, Nelson BJ. “Auto focusing in computer microscopy: selecting the optimal focus algorithm”. Microscopy Research and Technique, 65(3), 139–149, 2004.
18. [18] Yang G, Nelson B. “Wavelet-based autofocusing and unsupervised segmentation of microscopic images”. IEEE/RSJ International Conference on Intelligent Robots and Systems, Las Vegas, USA, 27-31 October 2003.
19. [19] Guyon I, Elisseef A. “An introduction to variable and feature selection”. Journal of Machine Learning Research, 3, 1157–1182, 2003.
20. [20] Jain S, Saha A. “Rank based univariate feature selection methods on machine learning classifiers for code smell detection”. Evolutionary Intelligence, 15, 609-638, 2022.
21. [21] Opoku Asare K, Terhorst Y, Vega J, Peltonen E, Lagerspetz E, Ferreira D. “Predicting depression from smartphone behavioral markers using machine learning methods, hyperparameter optimization, and feature importance analysis: exploratory study”. JIMR Mhealth Uhealth, 9(7), 1-17, 2021.
22. [22] Zhend A, Casari A. Feature Engineering for Machine Learning: Principles and Techniques for Data Scientists, 1st ed. USA, O'Reilly, 2018.
23. [23] Turan MK, Sehirli E. “A novel method to identify and grade DNA damage on comet images”. Computer Methods and Programs in Biomedicine, 147, 19–27, 2017.
24. [24] Ozdemir R, Turanli M. “Comparison of machine learning classification algorithms for purchasing forecast”. Journal of Life Economics, 8, 59–68, 2021.
25. [25] Güvenç E, Çetin GC, Koçak H. “Comparison of KNN and DNN classifiers performance in predicting mobile phone price ranges”. Advances in Artificial Intelligence Research, 1, 19–28, 2021.
26. [26] Akar Ö, Güngör O. “Rastgele orman algoritması kullanılarak çok bantlı görüntülerin sınıflandırılması”. Journal of Geodesy and Geoinformation, 1(2), 139–146, 2012.
27. [27] Ashour AS, Guo Y, Hawas AR. “Ensemble of subspace discriminant classifiers for schistosomal liver fibrosis staging in mice microscopic images”. Health Information Science and Systems, 6(1), 1-10, 2018.
28. [28] Fawcett T. “An introduction to ROC analysis”. Pattern Recognition Letters, 27(8), 861–874, 2006.
29. [29] Powers DMW. “Evaluation: from precision, recall and f-measure to ROC, informedness, markedness & correlation”. International Journal of Machine Learning Technologies, 2(1), 37–63, 2011.
30. [30] Kryl M, Danys L, Jaros R, Martinek R, Kodytek P, Bilik P. “Wood recognition and quality imaging inspection systems”. Journal of Sensors, 2020, 1-19, 2020.
31. [31] Martins J, Oliveira LS, Nisgoski S, Sabourin R. “A database for automatic classification of forest species”. Machine Vision and Applications, 24, 567–578, 2013.
32. [32] Yadav AR, Dewal ML, Anand RS, Gupta S. “Classification of hardwood species using ANN classifier”. Fourth National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics, Jodhpur, India, 18-21 December 2013.
33. [33] Filho PLP, Oliveira LS, Britto AS. “Forest species recognition using color-based features”. 20th International Conference on Pattern Recognition, Istanbul, Turkey, 23-26 August 2010.
34. [34] Yusof R, Rosli NR, Khalid M. “Using Gabor filters as image multiplier for tropical wood species recognition system”. 12th International Conference on Computer Modelling and Simulation, Cambridge, UK, 24-26 March 2010.
35. [35] Nasirzadeh M, Khazael AA, Khalid MB. “Woods recognition system based on local binary pattern”. 2nd International Conference on Computational Intelligence, Communication Systems and Networks, Liverpool, UK, 28-30 July 2010.
36. [36] Yusof R, Rosli NR, Khalid M. “Tropical wood species recognition based on Gabor filter”. 2nd International Congress on Image and Signal Processing, Tianjian, China, 17-19 October 2009.
37. [37] Tou JY, Tay YH, Lau PY. “A comparative study for texture classification techniques on wood species recognition problem”. Fifth International Conference on Natural Computation, Tianjian, China, 14-16 August 2009.
38. [38] Tou JY, Tay YH, Lau PY. “One-dimensional grey-level co-occurrence matrices for texture classification”, International Symposium on Information Technology, Kuala Lumpur, Malaysia, 26-28 August 2008.
39. [39] Khalid M, Lee ELY, Yusof R, Nadaraj M. “Design of an intelligent wood species recognition system”. International Journal of Simulation: Systems, Science and Technology, 9(3), 9–19, 2008.
40. [40] Tou JY, Lau PY, Tay YH. “Computer vision-based wood recognition system”. International Workshop on Advanced Image Technology, Bangkok, Thailand, 8-9 January 2007.