Search

In sports or fitness training, nonstandard movements will affect the training effect and even lead to sports injuries. However, the standard movements of various sports activities need professional guidance, so it is difficult to find out whether the movements are standard or not. In recent years, body pose estimation has become a hot topic in computer vision research. A deep learning model can effectively identify the human nodes and movement trajectory in pictures or videos and evaluate the movements of the target human body. However, the movement process is generally covered by others or the situation of nearby personnel, which leads to the deviation of the movement recognition of the human body and affects the evaluation of the movement. Thus, it is unable to effectively correct the wrong movement, but will mislead the training personnel. Therefore, this paper proposes a novel decision support model for sports training based on association rules. We use posterior probability settings to reveal the weights of the discriminative ability of attribute items and set the classification performance to reflect the weights of three measures to evaluate credit contribution. Thus, the learning threshold setting reflects the weight of the decision-making ability of sports training. Furthermore, compared with traditional association rules, attribute items, frequent item sets, and classification rules that can improve the decision-making performance of sports training are discovered, which complement the deficiencies of different measures. Finally, using the weighted voting strategy to fuse the decision-making information of the classification rules, we can effectively assist in sports training so that the coach can work out corresponding countermeasures and realize scientific management.

At present, the problem of pedestrian detection has attracted increasing attention in the field of computer vision. The faster regions with convolutional neural network features (Faster R-CNN) are regarded as one of the most important techniques for studying this problem. However, the detection capability of the model trained by faster R-CNN is susceptible to the diversity of pedestrians&rsquo, appearance and the light intensity in specific scenarios, such as in a subway, which can lead to the decline in recognition rate and the offset of target selection for pedestrians. In this paper, we propose the modified faster R-CNN method with automatic color enhancement (ACE), which can improve sample contrast by calculating the relative light and dark relationship to correct the final pixel value. In addition, a calibration method based on sample categories reduction is presented to accurately locate the target for detection. Then, we choose the faster R-CNN target detection framework on the experimental dataset. Finally, the effectiveness of this method is verified with the actual data sample collected from the subway passenger monitoring video.

Vehicle-mounted panoramic system is important safety assistant equipment for driving. However, traditional systems only render fixed top-down perspective view of limited view field, which may have potential safety hazard. In this paper, a texture mapping algorithm for 3D vehicle-mounted panoramic system is introduced, and an implementation of the algorithm utilizing OpenGL ES library based on Android smart platform is presented. Initial experiment results show that the proposed algorithm can render a good 3D panorama, and has the ability to change view point freely.

Recently, Scale Invariant Feature Transform (SIFT) algorithm is widely used in feature extraction and image matching. However, it has some defects, such as large volume of computational data and low efficiency of image matching. To address these defects, adaptive feature extraction and image matching based on Haar Wavelet Transform and SIFT (AHWT-SIFT) is proposed in this paper. In view of the characteristics of Haar wavelet, the low-frequency components of image can be decomposed adaptively by DWT, which represents the main features of the image and avoids the high-frequency of instability redundant information. Then SIFT is applied in these low-frequency components to extract the feature points. Furthermore, nearest neighbor algorithm is utilized for image matching. The experimental results have shown that the proposed scheme not only retains the general characteristics of SIFT, but the speed and accuracy of feature points matching have been greatly improved.

In this paper, a novel method for fabricating a static induction thyristor has been put forward, using silicon direct bonding instead of traditional epitaxy during the construction of a cathode. Thus, an obvious improvement of the breakdown value of gate–cathode junction has been observed and consequently the gate controllability on anode voltage has been enhanced. The bonded interface has been studied. Some adjustments in technology have been adopted to enhance the bonding quality. A way to guarantee the consistency of the breakdown voltage of gate junction with respect to the cathode has been advanced. Some measurements of I–V characteristic of SDB–SITH have been carried out and the practical result is also listed. A comparison between the SDB–SITH and epitaxial SITH has been made, mainly on I–V and the switching time.

Single lithium battery protection circuit is easy to implement, but it is not widely used at present. This paper introduces a method of realizing power management for a lithium ion battery pack of multi-cell in series. A chip named S-8261 is the main component of the circuit system. The chip was designed for single-cell application. But we use it in 10 series battery protection system. The circuit system can provide over charge, over discharge, over current and short circuit protection for lithium battery. Circuit uses a simple and effective measure for energy balance which can improve the discharge efficiency of lithium batteries. At the same time there is a way to solve the heating phenomenon occurred during using. The system is simple to implement, low cost, and also efficient.

2D barcode identification is a hot topic in the field of image processing. This paper shows the analysis of 2D barcode structure and the development at home and abroad. A 2D barcode recognition system based on image processing was proposed. The segmentation and denoising algorithm were improved. The identification and decoding of 2D barcode was realized in VC + + programming environment. The recognition efficiency of the barcode was successfully increased. It pushes the promotion and development of 2D barcode.

This paper aims to explore a method about electrocardiogram (ECG) signal denoising based on Hilbert-Huang Transform. The empirical mode decomposition method can decompose the noisy signal into a number of Intrinsic Mode Functions. Energy analysis is conducted on the IMFs to find out the boundary between the noisedominated IMFs and ECG signal dominated IMFs accurately. The most noisy IMFs are denoised by using Donoho soft-threshold denoising method. The denoised high frequency IMFs are added to the low frequency IMFs to reconstruct the original signal. The simulation experiments show that this method is simpler than the wavelet denoising method. It is not necessary to choose wavelet basis or determine the number of layers and the threshold. The proposed method can come close to or achieve the best level of wavelet denoising.

This paper describes a teaching method by Pspice simulation in the course of electric circuits. Pspice's rich and powerful functions provide various methods to analyze circuits. Pspice assisted teaching method can be used at different levels to help students solve problems quickly and efficiently, or help them understand the second-order circuit better, even apply this way to some more complex circuits.

Joint source-channel decoding (JSCD) techniques have been proposed to reduce symbol error rate and bit error rate in decoding a sequence through exploiting the residual redundancy, but these techniques always employ only spatial residual redundancy. In this paper, a new JSCD scheme which exploits both spatial and temporal redundancy in motion vector data (MVD) is proposed. Simulation results show that our proposed scheme leads to obvious improvement over the scheme without using the a priori information of temporal redundancy.

This paper introduces a new I/sub DDT/ test generation method based on the fact that a gate delay may differ from its nominal value. A Bayesian optimization algorithm based on genetic algorithms is utilized for I/sub DDT/ test generation. The paper proposes a fitness function to evaluate the evolution of the test patterns and uses a pseudo-probability method to create the first generation test patterns. The test patterns generated by the new approach are proved to be valid by a waveform simulator based on Boolean processes, even though gate delays are assigned randomly from 50% to 150% of their nominal values, and the number of the evolution generation can be reduced about 25%.

One of the key technologies for optical fiber vibration pre-warning system (OFVWS) refers to identifying the vibration source accurately from the detected vibration signals. Because of many kinds of vibration sources and complex geological structures, the implement of identifying vibration sources presents some interesting challenges which need to be overcome in order to achieve acceptable performance. This paper mainly conducts on the time domain and frequency domain analysis of the vibration signals detected by the OFVWS and establishes attribute feature models including an energy information entropy model to identify raindrop vibration source and a fundamental frequency model to distinguish the construction machine and train or car passing by. Test results show that the design and selection of the feature model are reasonable, and the rate of identification is good.