Search

This paper describes a study on speaker recognition using the ECAPA-TDNN architecture, which stands for Extended Context-Aware Parallel Aggregations Time-Delay Neural Network. It utilizes X-vectors, a method for extracting speaker features by converting speech into fixed-length vectors, and introduces a squeeze-and-excitation block to model dependencies between channels. In order to better explore temporal relationships in the context of speaker recognition and improve the algorithm's generalization performance in complex acoustic scenarios, this study adds input gates and forget gates to the ECAPA-TDNN architecture, combining them with CIFG (Convolutional LSTM with Input and Forget Gates) modules. These are embedded into a residual structure of multi-layer aggregated features. A sub-center Arcface, an improved loss function based on Arcface, is used for selecting sub-centers for subclass discrimination, retaining advantageous sub-centers to enhance intra-class compactness and strengthen the robustness of the network. Experimental results demonstrate that the improved ECAPA-TDNN-CIFG in this study outperforms the baseline model, yielding more accurate and efficient recognition results. [ABSTRACT FROM AUTHOR]

Taking Uyghur character recognition as an example, this paper use a method of image segmentation which combines traditional methods with CNN, and transforms and implements the Uyghur character on intelligent devices. Firstly, after analyzing the application of several general segmentation methods, this paper finds some shortcomings but also some ideas for Uighur segmentation. Then, starting from the characteristics of Uyghur language, such as structure, word formation and input habits, the author studies the idea of Uyghur adhesive language segmentation from the perspective of language characteristics, puts forward the basic algorithm of Uyghur symbol segmentation, and applies the Uyghur character adhesion segmentation based on minimum spanning tree and multi-queue primitive merging model to improve the segmentation efficiency. In addition, in order to solve the limitations of the traditional handwriting recognition framework of "preprocessing + feature extraction + classifier", this paper puts forward a new solution of Uyghur handwriting recognition technology combining prior domain knowledge with CNN, constructs a Uyghur handwriting recognition model on the basis of CNN and random elastic deformation, and also improves the recognition rate of the system by combining domain knowledge with CNN. According to the experimental results, we can conclude that the mixed method proposed in this paper can effectively break the technical bottleneck of traditional methods, thus improving the efficiency of segmentation and recognition. [ABSTRACT FROM AUTHOR]

The stator winding inter-turn short fault of permanent magnet synchronous motor (PMSM) is a common motor fault with high probability. If this fault is not detected and handled in time, small initial fault will rapidly develop into other faults such as grounding short, inter-phase short, winding short et al., and even the motor damage. In order to diagnose the inter-turn short fault of PMSM, this paper firstly builds the model of the inter-turn short fault of PMSM, and then proposes a fast and simple fault diagnosis method for extracting the features of the inter-turn short fault under noise. For the first time, by introducing the full-short-circuit resistance R f that relates with the fault severity factor δ and by estimating the inductance of inter-turn short fault motor on-line according to the three principles of the inductance calculation, which is used in the inter-turn short fault of transformer winding, the inter-turn short analysis model of PMSM is deduced. The model can be used to simulate the performance not only the health of PMSM but also the fault of inter-turn short fault of PMSM in one phase winding. This paper analyses the transient and steady state performance of 3-phase unbalanced current, speed and electromagnetic torque ripple and the relationship between full-short-circuit resistance and short current under fault condition. Finally, a fast and simple on-line fault diagnosis method of PMSM with inter-turn short fault is proposed based on phase sensitive detection (PSD) algorithm. The method uses the fundamental and third harmonic features of inter-turn short fault current. The simulation results show that the model can be well suited for the simulation of PMSM with inter-turn short fault, and the method can effectively extract the features and diagnose the fault. [ABSTRACT FROM AUTHOR]

Facial recognition is a technique for biometric identification, generally applied to finance, cybersecurity, public security, property management, and other important management departments. In this paper, the facial recognition system based on improved LBP and PCA mathematical algorithm is used to extract the relationship between the facial centre point and other phase domain points. After using PCA to reduce the dimension, the facial features are obtained. The facial recognition system designed by the thesis covers face image creation, face image pre-processing, face image recognition and result output module. Studies have shown that the system's face recognition accuracy can be as high as 90%. [ABSTRACT FROM AUTHOR]

An improved method based on immune genetic, which is successfully used to fuse the infrared and visible images,is proposed in this paper. Feature vectors are constructed by extracting the characteristics of infrared and visible images, and then apply the modified immune genetic algorithm to fuse the image.The experimental results show that the improved algorithm has a good search ability and high convergence speed. Meanwhile, the method can also reduce the vector feature dimensions effectively on the basis of maintaining sufficiently effective information , the efficiency of target recognition is improved significantly. [ABSTRACT FROM AUTHOR]

This paper studies pedestrian detection algorithm based on trapezoidal feature of local model, and applied to histogram intersection kernel support vector machine (HIKSVM) for verification. In comparison with traditional asymmetry, rectangle and triangle features, the experimental results from trapezoid feature of local model indicate that it can more effectively describe the pedestrian's postures and promote the accuracy and robustness of pedestrian detection. Experimental results show that the proposed algorithm is robust and accurate against cluttered dynamical background, occlusion and the object deformation, and tested in many pedestrian datasets and achieved good results. [ABSTRACT FROM AUTHOR]