As an important driving device, the permanent magnet synchronous motor (PMSM) plays a critical role in modern industrial fields. Given the harsh working environment, research into accurate PMSM fault diagnosis methods is of practical significance. Time–frequency analysis captures the rich features of PMSM operating conditions, and convolutional neural networks (CNNs) offer excellent feature extraction capabilities. This study proposes an intelligent fault diagnosis method based on continuous wavelet transform (CWT) and CNNs. Initially, a mechanism analysis is conducted on the inter-turn short-circuit and demagnetization faults of PMSMs, identifying and displaying the key feature frequency range in a time–frequency format. Subsequently, a CNN model is developed to extract and classify these time–frequency images. The feature extraction and diagnosis results are visualized with t-distributed stochastic neighbor embedding (t-SNE). The results demonstrate that our method achieves an accuracy rate of over 98.6% for inter-turn short-circuit and demagnetization faults in PMSMs of various severities.

With the acceleration of global population aging, the incidence of knee osteoarthritis (KOA) has risen significantly, placing unprecedented pressure on healthcare resources and creating an urgent need for automated segmentation technologies to enhance clinical diagnostic efficiency. Therefore, this paper proposes a novel multi-target tissue segmentation network for knee joints, SwMrNet, which integrates improved Swin Transformer units and a proposed multi-scale residual module within the decoder to enhance both segmentation accuracy and robustness. Firstly, a sliding-window mechanism is used to iteratively exchange feature information, allowing for the extraction of global tissue features. Then, features are extracted at multiple scales, with residual connections preserving the fine details of each tissue type. Through the repeated fusion of global and local features, the SwMrNet segmentation performance and robustness are significantly enhanced. Finally, the proposed model was evaluated on a public knee MRI dataset and a local clinical knee MRI dataset. On the public dataset, the model achieved a Dice score of 98.2%, with Dice scores for all segmented tissues exceeding 94%. On the local clinical dataset, the model showed visually consistent segmentation results, suggesting its potential as an efficient multi-tissue segmentation tool for automated knee joint analysis and auxiliary clinical assessment.