Search

B-mode ultrasound and ultrasound elastography are two routine diagnostic modalities for breast cancer. Unfortunately, few efforts have paid attention to learn bi-modal ultrasound jointly. By combining multi-view deep mapping-based feature representation with SVM-based classification, we proposed a novel integrated deep learning model, multi-view deep neural network support vector machine (MDNNSVM), to achieve breast cancer diagnosis on bi-modal ultrasound. In particular, multi-view representation learning extracts and fuses the various ultrasound characteristics (also including hardness information of soft tissue) effectively to differentiate benign breast lesions from malignant. Further, the SVM-based objective function is used to learn a classifier jointly with DNN to improve diagnostic accuracy significantly. The experimental results on a real-world dataset of breast cancer verify the effectiveness of the MDNNSVM with the best value of classification accuracy (86.36%) and AUC (0.9079).

Aiming at the typical underwater search and rescue task of AUV swarm, this paper establishes an integrated AUV swarm model on underwater target characteristics, underwater acoustic and optical equipment performance, underwater environment, etc. The task scheduling is based on the integrated AUV swarm model. Moreover, this paper proposes a construction method of AUV swarm task scheduling system by combineing task scheduling of AUV swarm with DoDAF architecture framework. In addition, the UPDM is used to verify and analyze the proposed method, and the visiuable task scheduling system is executed and displays in 3D video form. It is shown that the integrated modeling of the AUV swarm is all-round and necessary, and the UPDM provides a suitable simulation platform and efficiency evaluation tool for AUV swarm task scheduling.

B-mode ultrasound (BUS) imaging is widely used for diagnosis of breast tumors. In recent years, another ultrasound modality, namely ultrasound elastography (UE), has also been successfully applied in clinical practice. Although the combination of bimodal ultrasound imaging can improve the diagnosis accuracy for breast tumors, the single-modal ultrasound-based diagnosis is more popular in clinical practice, especially in the rural areas. Since transfer learning (TL) can transfer knowledge from a source domain to a target domain to help train more effective model, we propose to develop a single-modal ultrasound-based computer-aided diagnosis (CAD) with the transferred information from an additional modality in this work. We propose a projective model (PM) based multilayer kernel extreme learning machine (ML-KELM-PM) algorithm, which performs the parameter transfer approach in classifier to perform TL in CAD. The experimental results on a bimodal ultrasound image dataset show that the proposed ML-KELM-PM algorithm outperforms all the compared algorithms for the single-modal ultrasound-based CAD for breast tumors.

This paper introduces the grid nesting techniques and their application on the sand-dust storm forecasting model CUACE-Dust (CMA unified Atmosphere Chemistry Environment-Dust), which is coupled with PSU/NCAR mesoscale model. Based on CUACE-Dust, three numerical experiments are designed to show the advantage of the grid nesting techniques. In the key area that attracts attention, the nesting techniques help us get a result as good as that of the finer grid for the total area, but its time cost has been decreased to about 1/6

Through the research of data integrated technology of the CAD system in networked manufacturing environment, this article discusses how to use XML to describe and define the data information of the system, how to solve the problems of getting the part information, generating the BOM list and the mapping and data conversion between parts BOM and XML document, designs the data integration framework of the valve CAD system, realizes the data transmission of the CAD system and other systems based on CORBA platform.

Based on the three-dimensional Liu chaotic system, a state variable was added to construct a new hyperchaos Liu system. Through numerical simulation, the Lyapunov index spectrum, bifurcation diagram and phase diagram of the new hyperchaos Liu system were studied. Furthermore, its hardware circuits were implemented, and the corresponding waves of chaos were observed by oscillograph. At the same time, by employing the synchronization control method of single variable coupling feedback, the synchronous circuits of the two new hyperchaos Liu system were designed. In addition, the synchronous control was studied via experiments, and the parameter range for controlling the synchronization was found.

Based on WCF technology study, the 3D geographic information services was realized, using the character of WCF of supporting various distributed technologies. The distributed 3D geographic information system which is applicable in different networks and applications was built.

To solve the problem of product data exchange and sharing among heterogeneous systems, the key technologies were discussed, which included manufacturing process information modeling, file conversion from STEP AP203/AP214 to AP224 with the conversion ware of STEP neutral files. CAD systems must output product data with STEP standard. CAPP systems obtained geometry and manufacturing process information of the parts with STEP neutral files, and the corresponding process design was put up. Then By translating the STEP files into corresponding XML neutral files with the criterions constituted by the STEP AP28, XML schema library is setup. Thus, product data exchange and sharing of manufacturing information can be realized.

The classification of the uterine myoma and adenomyosis from their ultrasound images mainly depends on doctors' experience and lacks objective criterions. Here a novel classification method is proposed using the multiresolution analysis and the orientational fractal analysis. Firstly, texture features under various resolutions and orientational fractal features are obtained from ultrasound images. Then the feature selection (FS) process is implemented using the sequential forward selection algorithm (SFS). Finally a classifier based on the support vector machine (SVM) is set up to classify the images into normal (Nor), myoma (Myo) and adenomyosis (Ad) cases. From the application of 27 Nor, 45 Ad and 74 Myo cases, it is showed that orientational fractal features and some multiresolution texture features are sensitive to the uterine Myo and Ad classification. The SVM classifier with the selected features may be useful in the practical classification.

The geometric sequences have the property of ideal r-tuple distribution and controllable complexity, and Bent sequences have good correlation properties. The construction of a Bent sequence generator with feedforward taps is presented. >

The paper proposes a dynamic recognition neural network model (DRNN), which has the advantages of some neural networks, so it is learns fast, calculates simply, and constructs flexibly. The model is suitable for dynamic recognition, such as speech recognition. Experiments on the Chinese isolated digital recognition were carried out. The results show that the model has potential. >

A class of pseudorandom sequences over GF(p/sup m/) which are the generalization of m-sequences are constructed, and the periodic auto- and cross-correlation functions of the generalized m-sequences are analyzed. The cross-correlation function of two generalized m-sequences over GF(p/sup m/) with period p/sup mn/-1 is calculated in terms of the cross-correlation of two smaller m-sequences over GF(p) with period p/sup m/-1. Classes of sequences which have small peak values of cross-correlations are obtained. >