Tiny defect detection is a knotty task in industrial electronics production. Existing traditional and deep learning methods have achieved satisfactory performance, however, they still face challenges in accuracy, generalization ability, and computational complexity. Therefore, this study designs a tiny defect detection-based you only look once (TDD-YOLO) model and proposes an innovative compression training strategy to train on low-resolution images and test on original images. First, a four-ME layers structure is adopted to the backbone network, to integrate more underlying information and extract effective features. In addition, a miniature detection head is incorporated into the head network to improve the accuracy and generalization performance of you only look once (YOLO). Meanwhile, TDD-YOLO introduces wise intersection over union (W-IoU) to reevaluate the loss of bounding box regression (BBR) and reduce false negatives by fitting the model well to regular quality anchor boxes. Finally, an image compression method at different ratios is applied in the proposed compression training strategy, to reduce computational complexity and surprisingly further improve accuracy. Comprehensive experiments on several variable compressed datasets which are based on a public printed circuit board (PCB) defect dataset validate the effectiveness of our theoretical approach and illustrate that our proposed method outperforms state-of-the-art methods.

Earthquake prediction (EQP) is an extremely difficult task, which has been overcome by adopting various technologies, with no further transformation so far. The negative selection algorithm (NSA) is an artificial intelligence method based on the biological immune system. It is widely used in anomaly detection due to its advantages of requiring little normal data to detect anomalies, including historical seismic-events-based EQP. However, NSA can suffer from the undesirable effect of data drift, resulting in outdated normal patterns learned from data. To tackle this problem, the data changes must be detected and processed, stimulating fast algorithmic adaptation strategies. This study proposes a dendritic cell algorithm (DCA)-based adaptive learning method for drift detection and negative selection algorithm (DC-NSA) that dynamically adapts to new input data. First, this study adopts the Gutenberg–Richter (GR) law and other earthquake distribution laws to preprocess input data. Then, the NSA is employed for EQP, and then, the dendritic cell algorithm (DCA) is employed to detect changes to trigger gradient descent strategies and update the self-set in NSA. Finally, the proposed approach is implemented to predict the earthquakes of MW > 5 in Sichuan and the surroundings during the next month. The experimental results demonstrate that our proposed DC-NSA is superior to the existing state-of-the-art EQP approaches.