1. Home
  2. Vol 4, No 3 (2023)
  3. Kurdthongmee

Comprehensive Evaluation of Deep Neural Network Architectures for Parawood Pith Estimation

Wattanapong Kurdthongmee

Abstract


Accurate pith estimation is crucial for maintaining the quality of wood products. This study delves into deep learning techniques for precise Parawood pith estimation, employing popular convolutional neural networks (ResNet50, MobileNet, and Xception) with adapted regression heads. Through variations in regression functions, optimizers, and training epochs, the most effective models were pinpointed. Xception, coupled with Huber Loss regression, Nadam optimizer, and 200 epochs, showcased superior performance, achieving a 4.48 mm mean error (with a standard deviation of 3.69 mm) in Parawood. Notably, benchmarking on the Douglas Fir dataset yielded similar results (2.81 mm mean error, standard deviation: 1.57 mm). These findings underscore deep learning's potential for Parawood and Douglas Fir pith estimation, offering substantial benefits to wood industry quality control and production efficiency. By harnessing advanced machine learning techniques, this study advances wood industry processes, promoting the adoption of state-of-the-art technology in forestry and wood science.

 

Doi: 10.28991/HIJ-2023-04-03-06

Full Text: PDF


Keywords


Wood Pith Detection; Parawood; Deep Learning; Resnet; Mobilenet; Xception; Image Augmentation; Regression; Accuracy.

References


Islam, M. N., Rahman, F., Das, A. K., & Hiziroglu, S. (2022). An overview of different types and potential of bio-based adhesives used for wood products. International Journal of Adhesion and Adhesives, 112, 102992. doi:10.1016/j.ijadhadh.2021.102992.

Hakkou, M., Pétrissans, M., Zoulalian, A., & Gérardin, P. (2005). Investigation of wood wettability changes during heat treatment on the basis of chemical analysis. Polymer Degradation and Stability, 89(1), 1–5. doi:10.1016/j.polymdegradstab.2004.10.017.

Denih, A., Putra, G. R., Kurniawan, Z., & Bahtiar, E. T. (2023). Developing a Model for Curve-Fitting a Tree Stem’s Cross-Sectional Shape and Sapwood–Heartwood Transition in a Polar Diagram System Using Nonlinear Regression. Forests, 14(6), 1102. doi:10.3390/f14061102.

Hänsel, A., Sandak, J., Sandak, A., Mai, J., & Niemz, P. (2022). Selected previous findings on the factors influencing the gluing quality of solid wood products in timber construction and possible developments: A review. Wood Material Science & Engineering, 17(3), 230-241. doi:10.1080/17480272.2021.1925963.

Longuetaud, F., Schraml, R., Mothe, F., Ravoajanahary, T., Decelle, R., Constant, T., ... & Uhl, A. (2023). Traceability and quality assessment of Norway spruce (Picea abies (L.) H. Karst.) logs: the TreeTrace_spruce database. Annals of Forest Science, 80(1), 9. doi:10.1186/s13595-023-01178-8.

Longuetaud, F., Pot, G., Mothe, F., Barthelemy, A., Decelle, R., Delconte, F., ... & Viguier, J. (2022). TreeTrace project database on traceability and quality assessment of round woods: Douglas fir sampling. Annals of Forest Science, 79(46), 1-22. doi:10.1186/s13595-022-01163-7.

Longuetaud, F., Leban, J. M., Mothe, F., Kerrien, E., & Berger, M. O. (2004). Automatic detection of pith on CT images of spruce logs. Computers and Electronics in Agriculture, 44(2), 107–119. doi:10.1016/j.compag.2004.03.005.

Entacher, K., Hegenbart, S., Kerschbaumer, J., Lenz, C., Planitzer, D., Seidel, M., Uhl, A., & Weiglmaier, R. (2008). Pith detection on CT-Cross-Section images of logs: An experimental comparison. 2008 3rd International Symposium on Communications, Control and Signal Processing. doi:10.1109/isccsp.2008.4537273.

Krähenbühl, A., Kerautret, B., Debled-Rennesson, I., Longuetaud, F., & Mothe, F. (2012). Knot Detection in X-Ray CT Images of Wood. Advances in Visual Computing. ISVC 2012. Lecture Notes in Computer Science, 7432, Springer, Berlin, Germany. doi:10.1007/978-3-642-33191-6_21.

Boukadida, H., Longuetaud, F., Colin, F., Freyburger, C., Constant, T., Leban, J. M., & Mothe, F. (2012). PithExtract: A robust algorithm for pith detection in computer tomography images of wood – Application to 125 logs from 17 tree species. Computers and Electronics in Agriculture, 85, 90–98. doi:10.1016/j.compag.2012.03.012.

Kurdthongmee, W., Suwannarat, K., Panyuen, P., & Sae-Ma, N. (2018). A Fast Algorithm to Approximate the Pith Location of Rubberwood Timber from a Normal Camera Image. 2018 15th International Joint Conference on Computer Science and Software Engineering (JCSSE), Nakhonpathom, Thailand. doi:10.1109/jcsse.2018.8457375.

Kurdthongmee, W., & Suwannarat, K. (2019). An efficient algorithm to estimate the pith location on an untreated end face image of a rubberwood log taken with a normal camera. European Journal of Wood and Wood Products, 77(5), 919–929. doi:10.1007/s00107-019-01433-8.

Hanning, T., Kickingereder, R., & Casasent, D. (2003). Determining the average annual ring width on the front side of lumber. Optical Measurement Systems for Industrial Inspection III, Munich, Germany. doi:10.1117/12.500200.

Österberg, P., Ihalainen, H., & Ritala, R. (2004). Method for analyzing and classifying wood quality through local 2d-spectrum of digital log end images. Proceedings of International Conference on Advanced Optical Diagnostics in Fluids, Solids and Combustion, VSJ-SPIE'04, 4-6 December, 2004, Tokyo, Japan.

Norell, K., & Borgefors, G. (2008). Estimation of pith position in untreated log ends in sawmill environments. Computers and Electronics in Agriculture, 63(2), 155–167. doi:10.1016/j.compag.2008.02.006.

Kurdthongmee, W., & Suwannarat, K. (2019). Locating Wood Pith in a Wood Stem Cross Sectional Image Using YOLO Object Detection. 2019 International Conference on Technologies and Applications of Artificial Intelligence (TAAI), Kaohsiung, Taiwan. doi:10.1109/taai48200.2019.8959823.

Kurdthongmee, W. (2020). A comparative study of the effectiveness of using popular DNN object detection algorithms for pith detection in cross-sectional images of parawood. Heliyon, 6(2), 3480. doi:10.1016/j.heliyon.2020.e03480.

Kurdthongmee, W., Suwannarat, K., & Kiplagat, J. (2022). A Framework to Create a Deep Learning Detector from a Small Dataset: A Case of Parawood Pith Estimation. Emerging Science Journal, 7(1), 245–255. doi:10.28991/esj-2023-07-01-017.

He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, United States. doi:10.1109/cvpr.2016.90.

Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., ... & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint. doi:10.48550/arxiv.1704.04861.

Chollet, F. (2017). Xception: Deep Learning with Depthwise Separable Convolutions. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). doi:10.1109/cvpr.2017.195.

Decelle, R., Ngo, P., Debled-Rennesson, I., Mothe, F., & Longuetaud, F. (2022). Ant Colony Optimization for Estimating Pith Position on Images of Tree Log Ends. Image Processing on Line, 12, 558–581. doi:10.5201/ipol.2022.338.


Full Text: PDF

DOI: 10.28991/HIJ-2023-04-03-06

Refbacks

  • There are currently no refbacks.


Copyright (c) 2023 Wattanapong Kurdthongmee