We present a novel deep learning approach to extract point‐wise descriptors directly on 3D shapes by introducing Siamese Point Networks, which contain a global shape constraint module and a feature transformation operator. Such geometric descriptor can be used in a variety of shape analysis problems such as 3D shape dense correspondence, key point matching and shape‐to‐scan matching. The descriptor is produced by a hierarchical encoder–decoder architecture that is trained to map geometrically and semantically similar points close to one another in descriptor space. Benefiting from the additional shape contrastive constraint and the hierarchical local operator, the learned descriptor is highly aware of both the global context and local context. In addition, a feature transformation operation is introduced in the end of our networks to transform the point features to a compact descriptor space. The feature transformation can make the descriptors extracted by our networks unaffected by geometric differences in shapes. Finally, an N‐tuple loss is used to train all the point descriptors on a complete 3D shape simultaneously to obtain point‐wise descriptors. The proposed Siamese Point Networks are robust to many types of perturbations such as the Gaussian noise and partial scan. In addition, we demonstrate that our approach improves state‐of‐the‐art results on the BHCP benchmark.

To address the problems of low feature extraction accuracy, large bias of human motion pose recognition and posture recognition error, poor recognition effect, and low recognition rate of traditional human motion posture fast recognition algorithm, we propose a human motion posture fast recognition algorithm using multimodal bioinformation fusion. First, wavelet packet decomposition with sample entropy is used to extract the human motion posture hand features such as kurtosis, time domain feature skewness, and frequency domain feature electromyogram (EMG) integral value and time domain features such as mean, standard deviation, and interquartile distance of leg motion amplitude. Second, after normalizing the two features, the human hand and leg motion feature set is obtained, and finally the feature set is used to construct a human motion posture fast recognition model based on multimodal bioinformation fusion, and the feature set is input into the recognition model, which completes the fusion of human motion posture information by improving the typical correlation analysis method, and the fusion result is used as the input of the minimum distance classifier to achieve human motion posture fast recognition. The results show that the proposed algorithm has high accuracy of feature extraction, small bias of human motion posture recognition, the posture recognition error is -0.21∼0.02, the recognition rate is always above 95%, and the practical application effect is good.

With the advent of the computer network era, people like to think in deeper ways and methods. In addition, the power information network is facing the problem of information leakage. The research of power information network intrusion detection is helpful to prevent the intrusion and attack of bad factors, ensure the safety of information, and protect state secrets and personal privacy. In this paper, through the NRIDS model and network data analysis method, based on deep learning and cloud computing, the demand analysis of the real-time intrusion detection system for the power information network is carried out. The advantages and disadvantages of this kind of message capture mechanism are compared, and then a high-speed article capture mechanism is designed based on the DPDK research. Since cloud computing and power information networks are the most commonly used tools and ways for us to obtain information in our daily lives, our lives will be difficult to carry out without cloud computing and power information networks, so we must do a good job to ensure the security of network information network intrusion detection and defense measures.

No abstract