The manual Kellgren-Lawrence (KL) grading of osteoarthritis (OA) X-ray images is prone to inter-rater variability owing to the subjective nature of the evaluation process. In this study, we constructed an automatic KL grading system to realize intelligent assessment of OA severity in X-ray images through preprocessing, segmentation, feature extraction, and classification. The primary steps are as follows: First, preprocessing: This step integrates wavelet denoising, median filtering, and Medical_HED (MED-HED) edge detection to enhance image contrast and extract contours of key knee joint regions. Second, regional segmentation: The region of interest (ROI) is intercepted based on joint space localization, and the images of key regions of the knee joint are obtained by combining the connected region labeling method. Third, feature extraction: Fourier descriptors are employed to convert joint contours into shape vectors, facilitating extraction of geometric features that reflect OA severity. Fourth, grading classification: A template matching framework is constructed to compare shape vector similarity between samples and KL-grade templates, enabling automatic grading. Experimental results show that the system achieves an overall grading accuracy of 92.3% in 400 knee X-ray tests, outperforming conventional operators such as Canny and LOG. The proposed recognition method mitigates subjectivity in manual evaluation through multi-step image processing and shape analysis. Its high accuracy and efficiency make it a reliable computer-aided decision-making tool for clinical OA diagnosis, with potential to replace manual grading.

This paper proposes a NARX network-based modeling method to improve the real-time performance of dynamic component-level engine models. This method combines the component-level model with the NARX network black-box modeling technique. It collects training data by solving the common working equation of the component-level model (N-R model) using the Newton–Raphson method and uses a multilayer feedforward neural network to approximate the nonlinear mapping from the input state space to the solution of the common working equation. This approach eliminates the need for iterative computations in solving the common working equation, thus enhancing the real-time performance of the component-level model. Using the N-R model as the benchmark, simulation results show that the NARX network model with embedded component-level modeling has a maximum dynamic error of 0.061% during ground state modeling and 0.36% across the full envelope; the average runtime for full envelope modeling is 0.046 seconds, which is only 4.3% of that of the benchmark model; for step inputs, the component residual during the dynamic process is only 0.02% of that of the benchmark model. These findings demonstrate that the engine model established by this method has high accuracy, good real-time performance, and robustness.

Over the past few decades, several methods for decomposing electromyographic (EMG) signals into motor unit action potentials (MUAPs) have been reported, but there is still much room for improvement. Here, a new method for decomposing high-density surface EMG (HD-sEMG) signals is proposed. First, the firing time is extracted based on the linear minimum mean square error (LMMSE) algorithm. Second, the waveforms corresponding to the firing time are converted into images, and the images are spliced on the electrodes of the column in which the motor unit is located. Third, the images corresponding to the multiple firing times are clustered and averaged, and the motor unit action potential (MUAP) train is obtained iteratively. Finally, the extracted MUAPs are removed from the original signals, and the cycle further decomposes the residual signals. The simulation results show that the interval pulse trains can be reconstructed with a true positive rate (TPR) of greater than 91.32 ± 4.17%. According to the experimental results, 14-20 MUs were extracted, and the ratio of common discharges was determined to be 0.81 ± 0.03% via the two-source method. The simulation and experimental results show that the proposed method is accurate and efficient with potential applications in rehabilitation and motion control systems.

In typical image inpainting tasks, the locations and shapes of damaged or masked areas are often random and irregular. Vanilla convolutions, commonly employed in learning-based inpainting models, treat all spatial features as valid and share parameters across different regions. This approach can struggle with irregular damage patterns, leading to inpainted results that may suffer from color discrepancies and blurriness. In this paper, we introduce a novel operator known as Semi-Independent Convolution (SIConv) to tackle this challenge. The proposed SIConv, on top of the regular convolution with shared weights, also introduces dynamic terms that assign their own independent weights to each part of the image, and the overall computation is formulated as a shared convolution parameter with an additional term to describe the local structure. Qualitative and quantitative experiments demonstrate that our method outperforms the state-of-the-art, yielding clearer, more coherent, and visually convincing inpainting results.