Radhwan A. A. Saleh,Mehmet Zeki Konyar,Kaplan Kaplan,H. Metin Ertunç
Abstract:Abstract Visual inspection of defective tires post-production is vital for human safety, as faulty tires can lead to explosions, accidents, and loss of life. With the advancement of technology, transfer learning (TL) plays an influential role in many computer vision applications, including the tire defect detection problem. However, automatic tire defect detection is difficult for two reasons. The first is the presence of complex anisotropic multi-textured rubber layers. Second, there is no standard tire X-ray image dataset to use for defect detection. In this study, a TL-based tire defect detection model is proposed using a new dataset from a global tire company. First, we collected and labeled the dataset consisting of 3366 X-ray images of faulty tires and 20,000 images of qualified tires. Although the dataset covers 15 types of defects arising from different design patterns, our primary focus is on binary classification to detect the presence or absence of defects. This challenging dataset was split into 70, 15, and 15% for training, validation, and testing, respectively. Then, nine common pre-trained models were fine-tuned, trained, and tested on the proposed dataset. These models are Xception, InceptionV3, VGG16, VGG19, ResNet50, ResNet152V2, DenseNet121, InceptionResNetV2, and MobileNetV2. The results show that the fine-tuned VGG19, DenseNet21 and InceptionNet models achieve compatible results with the literature. Moreover, the Xception model outperformed the compared TL models and literature methods in terms of recall, precision, accuracy, and F1 score. Moreover, it achieved on the testing dataset 73.7, 88, 80.2, and 94.75% of recall, precision, F1 score, and accuracy, respectively, and on the validation dataset 73.3, 90.24, 80.9, and 95% of recall, precision, F1 score, and accuracy, respectively.
What problem does this paper attempt to address?
The paper attempts to address two main issues in tire defect detection:
1. **The presence of complex multi-textured rubber layers**: Tires are composed of complex, anisotropic multi-textured rubber layers, making automatic tire defect detection very challenging.
2. **Lack of a standard tire X-ray image dataset**: In the field of tire defect detection, there is no standard dataset available for training and testing models.
To solve these problems, the authors propose a tire defect detection model based on Transfer Learning (TL) and use a new dataset. This dataset contains 3,366 X-ray images of faulty tires and 20,000 images of qualified tires. The dataset covers 15 different defect types under different design patterns, but the main focus of the study is on the binary classification task, i.e., detecting whether a tire has a defect.
By fine-tuning, training, and testing nine common pre-trained models (Xception, InceptionV3, VGG16, VGG19, ResNet50, ResNet152V2, DenseNet121, InceptionResNetV2, and MobileNetV2) on the new dataset, the authors evaluated the performance of these models on the tire defect detection task. Experimental results show that the fine-tuned VGG19, DenseNet121, and InceptionNet models exhibit comparable results in the literature, while the Xception model outperforms other models in terms of recall, precision, accuracy, and F1 score. Specifically, the Xception model achieved a recall of 73.7%, a precision of 88%, an F1 score of 80.2%, and an accuracy of 94.75% on the test set, while on the validation set, it achieved a recall of 73.3%, a precision of 90.24%, an F1 score of 80.9%, and an accuracy of 95%.
In summary, the paper aims to improve the accuracy and efficiency of tire defect detection through transfer learning techniques using a new tire X-ray image dataset.