Wood defect identification using convolutional neural network features with support vector machine classifier
Accurate classification of wood surface defects is essential for maintaining product quality and minimizing material waste in the timber industry. However, achieving high classification accuracy is challenging due to the limited availability of labeled datasets, particularly across diverse wood spec...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2025
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/29384/ |
| Abstract | Abstract here |
| Summary: | Accurate classification of wood surface defects is essential for maintaining product quality and minimizing material waste in the timber industry. However, achieving high classification accuracy is challenging due to the limited availability of labeled datasets, particularly across diverse wood species. This study proposes a Convolutional Neural Network–Support Vector Machine (CNN-SVM) approach that leverages transfer learning for feature extraction and multi-class wood defect classification. Pre-trained CNN models were employed to extract discriminative features from wood surface images, which were then classified using a Support Vector Machine to enhance accuracy across nine defect classes. The method was evaluated based on classification accuracy and statistical validation. Among the tested models, the ResNet50 SVM combination demonstrated the most consistent and accurate performance. These findings suggest that the CNN-SVM approach offers a viable solution for improving automated wood defect classification. |
|---|