Additive Manufacturing: Stringing and Warping Detection Using MobileNet-SSD

Defect detection is an essential process in additive manufacturing, especially for 3D printing. This feature can potentially help in quality control, production efficiency, waste product and cost reduction. In this paper, we propose a deep learning-based approach using the MobileNet-SSD algorithm for detecting and differentiating faults on 3D artefacts during printing. The dataset was added after the initial training, allowing the model to be retrained to improve its accuracy. The research flow includes selecting the best deep learning algorithm for defect detection, data collection, creating a neural network model, image pre-processing, hyperparameter tuning, and model training and validation. The accuracy of the proposed model was evaluated, and the results achieved a mean average precision (mAP) of 28%. The proposed approach is effective in detecting defects compared to ResNet-SSD with mAP 21.4%.