125 results about "Cascade network" patented technology

A technique for use in gathering network activity-related information from cascaded network switches is provided. Using this technique, the information can be gathered without substantially reducing performance of the cascaded switches. In one embodiment, a single remote monitoring probe is connected via respective connections to each of the switches so as to receive the information from the switches. In another embodiment, only one of the switches is connected to the probe, and the other switches transmit their respective portions of the information to the switch connected to probe. The switch connected to the probe provides these portions of the information, as well as, any of its respective activity-related information to the probe. In this latter embodiment the switches may be connected by dedicated connections and switch ports that are used solely for communicating the activity-related information.

A QKD cascaded network (5) with loop-back capability is disclosed. The QKD system network includes a plurality of cascaded QKD relays (10, 20, 30) each having two QKD stations Alice (A) and Bob (B) therein. Each QKD relay also includes an optical switch (50). The optical switch is optically coupled to each QKD station in the relay, as well as to the input ports (PI) of the relay. In a first position, the optical switch allows for communication between adjacent relays. In a second position, the optical switch allows for pass-through communication between the QKD relays (10 and 30) that are adjacent the relay whose switch is in the first position. Also in the second position, the optical switch allows for communication between the QKD stations A and B within the relay. This, in turn, allows for diagnostic measurements to be made of one or both of the QKD stations via an optical path (90) that is entirely within the relay station enclosure (12, 22, 32).

The present invention discloses a cascade network building method based on the wavelength division multiplexing (xWDM) radio frequency remote unit, and mainly comprises the following steps: 1, the wavelength division multiplex (xWDM) is used in a radio frequency remote type BTS system for the cascade radio frequency remote unit RRU; 2, the wavelength division multiplex (xWDM) and the time division multiplex (TDM) are compounded and then used in the radio frequency remote type system, and different wavelengths are adopts to form the logical star connection between the BTS and different RRUs; 3, the wavelength division multiplex (xWDM) is used in the RRU of the cascade radio frequency remote type BTS system and in the base station system BTS; 4, the wavelength division multiplex (xWDM) is used in the cascade optical transmission system from the one-point main equipment to the multi-point auxiliary equipment. The network building method greatly saves the transmission resources of the system, enlarges the transmission capacity of the system, prevents the problem of gradual declination of the clocks which are restored and transmitted step by step in the TDM multiplexing mode, simplifies the problems in the estimation of time delay between the cascade systems, and is suitable for the base station system of multi-mode communication modes.

Embodiments of the present application disclose a medical image processing method, system and apparatus, and computer-readable storage medium. The method comprises the following steps: acquiring a sample image, wherein the sample image comprises an abnormal region, the abnormal region comprises at least one sub-abnormal region; Acquiring an image processing cascade network, the image processing cascade network comprising at least two levels of neural networks; Training the neural networks in the image processing cascade network with the sample images to obtain a fully trained image processingcascade network; Wherein the input of the first-level neural network is an unsegmented sample image, and the output of the last-level neural network is a segmentation result of the at least one sub-abnormal region; The trained image processing cascade network is used to process the image to be tested and determine the abnormal region in the image to be tested.

The invention provides a method for dispatching MCU resources, a business server and a system. The method comprises the following steps of: the business server determines a candidate MCU collection; the business server constructs a terminal list which records a terminal set needed by a current conference; each terminal set comprises at least one terminal; the business server distributes the terminal set to the MCU in the candidate MCU collection according to a scheduled strategy; the available resource supplied by the MCU distributed to the terminal end is larger than or equal to the sum of the distributed resources needed by the terminal end; according to the corresponding relationship between the terminal end and the MCU, the topological relation of a cascaded conference is determined. By applying the technical proposal, according to the available resources supplied by different MCU, the corresponding relationship between the MCU and the terminal set is determined; and according to the relationship, the topological relation of a cascaded network is established, thus realizing the automatic dispatching of the MCU resources in a TV conference system.

The invention discloses a stylized image generation method, device and equipment and a storage medium. The method comprises the following steps: carrying out style conversion on an original image through different styles of selected images to obtain a stylized original image; segmenting a target region from the stylized original image through a multi-task cascaded network according to a stylization target required by the original image; and fusing the segmented target region into the original image through an MRF model to generate a stylized image having the target region. The method can accurately classify the target region through the multi-task cascaded network and adopts the MRF model when the target region is fused with the original image, so that transition between the target regionand a background image of the original image can be smoother, the stylized image is allowed to take the content as contour and have a desired style, image stylization effect is ensured, and finally, specific region stylization or multi-region stylization of the image can be realized.

Through step of MAC address collision detection determines whether collision of MAC address happens between ports of network node at second layer of Ethernet. Through notifying step of MAC address collision control frame creates and sends a MAC address collision control frame. Through step of port control of MAC address collision carries out MAC address control for relevant ports. The invention also discloses device for controlling network MAC address collision including to more network ports, module for detecting MAC address collision, module for creating and sending control frame for MAC address collision, module for receiving the control frame for MAC address collision, and module for controlling MAC address collision. The invention also discloses system composed of cascaded network nodes at second layer for controlling network MAC address collision. The method, device and system better control MAC address collision in Ethernet.

A method for providing cascaded trie-based network packet search engines is provided. A search command is received at one of the network packet search engines. The search command comprises a specific search key. A determination of a longest prefix match based on the specific search key is made at the network packet search engine. A determination is made at the network packet search engine regarding whether the longest prefix match comprises an overall longest prefix match among the cascaded network packet search engines such that any of the cascaded network packet search engines may comprise the overall longest matching prefix independently of position relative to the other cascaded network packet search engines.

The invention discloses a posture estimation and human body analysis system based on multi-task deep learning. The system comprises a human body detection subnet and a posture estimation and human body analysis combined learning subnet. An input image firstly passes through a human body detection subnet to obtain information such as a human body position and a mask, and an interference-free single-person image is extracted from a multi-person image according to the information; the method further includes performing attitude estimation and human body analysis joint learning on the interference-free single-person image to obtain an attitude estimation result and a multi-granularity human body analysis result; and finally, combining the single-person posture estimation result and the multi-granularity human body analysis result to the original image. Different human body instances are distinguished based on human body postures, and a better human body detection effect is achieved on multi-person images; according to the invention, the accuracy of two tasks of posture estimation and human body analysis can be improved; and a cascade network structure is adopted for a human body analysis task, so that the human body analysis accuracy can be effectively improved, and finer analysis granularity expansion is facilitated.

The invention relates to a video saliency target detection method based on a cascade convolutional network and an optical flow, and the method comprises the steps: carrying out the pixel-level saliency prediction of an image of a current frame in the high scale, the middle scale and the low scale through employing a cascade network structure. A cascade network structure is trained by using an MSAR10K image data set, a saliency annotation graph is used as supervision information of training, and a loss function is a cross entropy loss function. after the training is ended, static saliency prediction is carried out on each frame of image in the video by using the trained cascade network. A classic Locus- Kanada algorithm is used to carry out optical flow field extraction. a three-layer convolutional network structure is used to construct a dynamic optimization network structure. the static detection result and the optical flow field detection result of each frame of image are spliced toobtain input data of the optimized network. And a Davis video data set is used to optimize the network, and pixel-level significance classification is carried out on the video frame by using a staticdetection result and optical flow information.

The invention discloses a base station system which includes a main control unit used for managing the base station system; a middle radiofrequency unit used for processing middle radiofrequency signals; a base band unit used for processing base band signals; a transmission unit which is connected with a network controller through a transmission interface; a model management unit used for allocating different working models of the base station system according to the networking forms of the base station system in a communication network; a first carrier frequency interface which is used for being communicated with a connected upper grade device during cascaded networks and a second carrier frequency interface which is used for being communicated with a connected lower grade device during the cascaded networks. The base station system which is provided by the embodiment of the invention can take part in the communication networking through working in different working models so as to support the flexible model of the communication networking. The base station system is convenient for operators to dilate, can reuse original equipment, avoids eliminating old base stations to reestablish new base stations and protects the investments of the operators.

The invention discloses a serialized face key point detection method with relay supervision based on deep learning, and the method comprises the following steps: giving a detected face image, designing a cascade network structure composed of a plurality of modules, and achieving the serialized prediction of key points through the structure through the relay supervision of each module; calculatingthe expectation of the key point coordinates on the output key point heat map by combining the soft maximum value function, completing the conversion from heat map regression to numerical value coordinate regression based on the key point coordinates. The respective advantages of the two regression methods are brought into full play. According to the method, on one hand, the defect that a heat mapregression method cannot perform end-to-end training is overcome, on the other hand, an additional data processing process is omitted, the model training process is accelerated, and the algorithm efficiency is improved. When numerical coordinate regression is carried out, a new segmented loss function is adopted, and finally high positioning accuracy is obtained.

The invention relates to a method for predicting protein association graphs on the basis of cascade neural network structures. The method as shown in an attached graph 1 includes steps of A, creating six subnets of neural networks and a cascade neural network; B, reading protein data sets and classifying the data sets according to protein lengths; C, carrying out training and learning on the subnets of the neural networks by the aid of back propagation algorithms; D, carrying out training and learning on the cascade neural network; E, predicting subnets of first-layer neural networks; F, predicting second-layer cascade neural networks to obtain the ultimate protein association graphs. The method has the advantages that the method is implemented by the aid of by the multiple neural networks, the cascade structures are formed, the protein association graphs are predicted, accordingly, the shortcoming that proteins with different lengths cannot be optimally treated by the aid of a method implemented by the aid of a single neural network can be overcome, and the prediction precision and stability can be improved; the method has an inherent concurrent characteristic, accordingly, the various subnets and the cascade network can be concurrently processed, and the computation efficiency can be improved.

The invention discloses a method for achieving face key point detection based on cascade MobileNet-V2 , which comprises the following steps: S1, obtaining a picture data set; S2, establishing a MobileNetV2 cascading network; S3, preliminarily determining the face key point, S4, cutting a face area, S5, determining the face area again, and S6, obtaining an accurate face key point. The invention adopts the MobileNet-V2 neural network with high speed and high precision, and improves it, and uses the two-stage neural network cascading method to improve the accuracy of face detection.; compared with an existing face key point detection model, the method has the advantages of being higher in speed and precision; Through a cascade mode, positioning of fine promotion and refinement is achieved, and cascade network training comprises the steps of firstly training a first-stage cascade network, achieving rough positioning of face key points, then cutting a face area, and then achieving accuratepositioning of the face key points through a second-stage cascade network.

The invention relates to the technical field of target attribute identification, and provides a small target vehicle attribute identification method based on feature fusion. The method comprises: firstly, constructing a small target vehicle attribute recognition network based on feature fusion, comprising a feature pyramid network, a regional nomination network and a small-size target cascade network; inputting a traffic image to be detected into the feature pyramid network, generating a feature map containing low-level edge detail information, middle-level stacking fusion scale information and high-level semantic information, and stacking and fusing the feature map to obtain a multi-scale feature map; inputting the multi-scale feature map into a regional naming network to generate a candidate box containing a target; inputting the multi-scale feature map and the candidate box into a small-size target positioning network at the same time, outputting target coordinate information, and cutting a target according to the information; and finally, inputting the sheared target into a small-size target classification network, and identifying and outputting the target and the category thereof. According to the invention, the accuracy of small-size target attribute identification can be improved, and the false identification rate and the missing identification rate are reduced.

The invention provides a networking method based on an unmanned aerial vehicle cluster. The networking method comprises unmanned aerial vehicles and routing relay nodes composed of the unmanned aerialvehicles, and the networking method comprises the following steps: planning a to-be-networked area within an activity range of the unmanned aerial vehicles according to a task, dividing the to-be-networked area into a plurality of subareas according to the wireless network coverage capability of each routing relay node composed of the unmanned aerial vehicles, selecting optimal hovering positionsof the unmanned aerial vehicles within each subarea, controlling the unmanned aerial vehicles to arrive at the hovering positions, forming a cascaded network communication routing system among the hovering unmanned aerial vehicles, allowing the access of a ground device, and completing the bidirectional data transmission with the ground device. By adoption of the networking method provided by theinvention, the conventional network communication of the ground device <---> unmanned aerial vehicles <---> a ground station central control unit <---> a public network is achieved, so that the real-time requirements similar to onsite rescue dispatching are met.

The invention relates to an automatic segmentation method for an abdominal CT liver lesion image based on a three-stage cascade network. The method comprises the following steps: S1, acquiring three-dimensional abdominal liver CT image data; S2, preprocessing and standardizing the obtained three-dimensional abdominal liver CT image data; S3, inputting the three-dimensional abdominal liver CT imagedata after preprocessing and data standardization into AuxResUnet liver image segmentation model, and then taking 3D maximum connected region from the obtained three-dimensional abdominal liver CT image data segmentation result to exclude false positive region, so as to obtain liver VOI; S4, adopting the segmentation result of the three-dimensional abdominal liver CT image data obtained by S3 asa mask of the CT liver image data to cover the liver VOI obtain by S3; S5, inputting the covered liver VOI into AuxResUnet liver image lesion segmentation model for lesion segmentation, and obtainingthe liver image lesion segmentation result. The image segmentation method provided by the invention can realize fast and accurate segmentation of liver and liver pathological changes.

The invention discloses a pulmonary nodule segmentation method based on a two-dimensional convolutional neural network. The pulmonary nodule segmentation method comprises the following steps: samplingpulmonary nodule positive and negative samples based on a weighted sampling strategy; training a two-dimensional convolutional neural network model according to the sampled data to obtain a trained two-dimensional convolutional neural network model; and predicting each voxel of the sample to be segmented by using the trained two-dimensional convolutional neural network model to obtain a pulmonarynodule segmentation result. When the sampling weight of non-pulmonary nodule voxels in a CT image is calculated, the grayscale information of non-pulmonary nodule tissues is considered, so that advanced features except the grayscale features are mined, and the heterogeneous property of pulmonary nodules is adapted; Fully sampling pulmonary nodules with different sizes by taking pulmonary nodule edge voxels as references; Local texture information and context information of the pulmonary nodules can be extracted based on the double-branch cascade network of the residual block; Through cascadeconnection of the two image blocks with different scales, segmentation of pulmonary nodules with small sizes is realized.

Apparatuses and methods of manufacturing same, systems, and methods for generating a convolutional neural network (CNN) are described. In one aspect, a minimal CNN having, e.g., three or more layers is trained. Cascade training may be performed on the trained CNN to insert one or more intermediate layers until a training error is less than a threshold. When cascade training is complete, cascade network trimming of the CNN output from the cascade training may be performed to improve computational efficiency. To further reduce network parameters, convolutional filters may be replaced with dilated convolutional filters with the same receptive field, followed by additional training/fine-tuning.

The invention discloses a cascaded residual neural network-based DOA and DOD joint estimation algorithm for a double-base co-prime MIMO array. The method comprises the steps of estimating the number of targets and estimating orientation angles of the targets. According to the method, a data processing part in traditional DOA and DOD estimation of double-base MIMO is improved. Compared with a traditional signal processing algorithm, a deep learning method has higher timeliness and better performance in the aspects of low signal-to-noise ratio, low snapshot, large orientation angle and robustness under coherent target conditions. The deep neural network adopted in the algorithm adopts a cascaded network structure. Firstly, a signal received by the array is subjected to DFT processing and then is subjected to correlation processing; the processed signal is input to the neural network to obtain DOA information of the signal; the DOA information is used as prior information to be input to the cascaded network to obtain DOD information of the signal; and finally, the DOA and DOD matching estimation of the signal is completed.

The invention provides a multi-person posture estimation method based on a deep cascade network and centroid differentiation coding. An estimation route from bottom to top is adopted; the problems ofjoint shielding and scale difference which cannot be solved by an existing algorithm are solved. A centroid differentiation code is designed as a correlation clue of the joint; a deep cascade networkbased on a two-way feature extraction module is established to complete extraction of joint point and centroid differentiation codes; and then a greedy reasoning strategy is proposed to realize robustmatching of joint points to multiple human body halves, finally spatial constraints are added between the halves, and human body splicing is completed by using a graph matching algorithm so as to realize rapid and efficient multi-person posture estimation.

The invention provides a node network monitoring method and device. The method comprises the steps that first, node parameter configuration is carried out; second, a monitored object is determined according to a user instruction; third, a corresponding monitoring module is called to monitor the monitored object according to the monitored object. The node network monitoring method and device can monitor various cascade network devices; in addition, the node network monitoring method and device can skip to network devices of downlink small nodes step by step to test downlink routers and the situation of VPN lines.