47results about How to "Suitable for practical application" patented technology

The invention provides a bidirectional quantum time synchronization method based on frequency entangled light sources. Two entangled light sources with same wavelength are placed at a place A and a place B respectively, and signal lights generated by the entangled light sources pass through an optical fiber circulator, are transmitted in opposite directions through a same optical fiber, and are detected by detectors at the opposite places; idle lights generated by the entangled light sources are detected by a detector at a uniform place of the entangled light sources; a clock a at the place arecords the time (tij} and time (t2j} for photons emitted by the entangled light source at the place a to reach the detectors at two places, and a clock b at the place b records time (t3j} and time (t4j} for photons emitted by the entangled light source at the place b to reach the detectors at two places, wherein j is representative of jth photon; the (tij} and (t2j} are subjected to cross-correlation operation to obtain t2-t1; the (t3j} and (t4j} are subjected to cross-correlation operation to obtain t4-t3; the clock difference t0 between the clock a and the clock b is equal to half of the difference of (t2-t1) and (t4-t3). The bidirectional quantum time synchronization method based on the frequency entangled light sources is high in measurement precision, and an applied device is simplein structure.

The invention discloses a malicious user discrimination method based on a self-organizing map neural network in a cognitive radio. The invention uses a self-organizing map (referred to as SOM) neural network to learn the distribution characteristics of an input energy matrix, and calculates the input amount according to the learning result. effective classification. First, the concept of "suspiciousness" is introduced, and its size is allocated according to the number of secondary users contained in each category after each training. In order to eliminate the defects of the traditional SOM neural network, the present invention further proposes the concept of "average suspiciousness". The specific steps include: obtaining the energy matrix, using the SOM neural network algorithm to train the energy matrix to obtain the classification matrix, calculating the "suspiciousness" of each secondary user, constructing the index matrix and repeating the training process, and calculating the "suspicious degree" obtained each time degree", that is, the "average suspicious degree", and use the "average suspicious degree" to classify secondary users to identify malicious users or normal users.

A coronary stenosis detection method, device, computer equipment, and storage medium provided by the application obtain detection results by inputting acquired images to be detected into a coronary stenosis detection model; wherein the coronary stenosis detection model includes a backbone network , a segmentation network and a stenosis analysis network, the outputs of the backbone network are respectively connected to the inputs of the segmentation network and the stenosis analysis network; the detection results include coronary artery segmentation results and coronary artery stenosis results. From the structure of the coronary stenosis detection model above, it can be seen that the coronary stenosis detection model includes a segmentation network and a stenosis analysis network. Therefore, the above coronary stenosis detection model is a multi-task model, which can simultaneously realize the detection of the input Segmenting the image, and performing stenosis detection on the input image to be detected, to obtain at least two types of detection results, namely coronary artery segmentation results and coronary artery stenosis results.