The existing face detection methods were affected by the network model structure used. Most of the face recognition methods had low recognition rate of face key point features due to many parameters and large amount of calculation. In order to improve the recognition accuracy and detection speed of face key points, a real-time face key point detection algorithm based on attention mechanism was proposed in this paper. Due to the multiscale characteristics of face key point features, the deep convolution network model was adopted, the attention module was added to the VGG network structure, the feature enhancement module and feature fusion module were combined to improve the shallow feature representation ability of VGG, and the cascade attention mechanism was used to improve the deep feature representation ability. Experiments showed that the proposed algorithm not only can effectively realize face key point recognition but also has better recognition accuracy and detection speed than other similar methods. This method can provide some theoretical basis and technical support for face detection in complex environment.

No abstract

Aiming at the existing problems with traditional integration technology, the system structure, service and workflow of OGSA-DAI are analyzed thoroughly. Based on OGSA-DAI extension, the data heterogeneity of semantic is shielded by schema mapping. An integration model is proposed to solve data integration and distributed query under the grid environment, achieving transparent access to distributed heterogeneous data and associated process.

In order to overcome the difficulty of extracting features from data and improve the accuracy of anomaly detection system, this paper proposes a novel anomaly detection method based on deep learning. We build a deep neural network model with multiple hidden layers to automatically learn features of data before detecting anomaly behaviors. The learned features from this network can enhance the discrimination of different behaviors. Moreover, an exactly sparse auto-encoder (ESAE) is proposed to achieve the pre-training of this network. This method does not require manual extraction of features, and is unsupervised, avoiding the difficulty of providing labeled data. Experimental results show that the proposed method could significantly improve the detection accuracy.

Abstract. Cloud services' QoS usually changes dynamically. Because of the significant influence caused by user's position and network conditions and work-loads, the reliability of the cloud service selection constitutes a great challenge. A cloud services multiple criteria selection model was proposed based on each period of QoS characteristics and user's personality preferences which could support dynamic QoS. First, attributes of available cloud service and user preferences were expressed by triangular fuzzy number in order to solve the uncertainty, then, decision matrix for each period of QoS was built based on the available cloud service, The decision matrix with the user preference weight were conducted decision fusion after user preferences weight was calculated, finally, The high grade cloud service sequence was obtained by using fuzzy TOPSIS.